Zero Waste within the food sector and an

evaluation of the package-free distribution

of two different food products
Vilhelm Sjölund
DIVISION OF PACKAGING LOGISTICS | DEPARTMENT OF DESIGN SCIENCES
FACULTY OF ENGINEERING LTH | LUND UNIVERSITY
2016

MASTER THESIS
 Zero Waste within the food sector
and an evaluation of the package-free
distribution of two different food
products
The environmental impact of different packaging models
of oats and rapeseed oil: a case study of a Danish
package-free retail store compared to a reference,
packaged scenario

Vilhelm Sjölund
Zero Waste within the food sector and an evaluation of
the package-free distribution of two different food
products
The environmental impact of different packaging models of oats and
rapeseed oil: a case study of a Danish package-free retail store compared to
a reference, packaged scenario.

Copyright © 2016 Vilhelm Sjölund

Published by
Department of Design Sciences
Faculty of Engineering LTH, Lund University
P.O. Box 118, SE-221 00 Lund, Sweden

Subject: Packaging Logistics (MTT920)

Supervisor: Katrin Molina-Besch
Co-supervisor: Karin Östergren, RISE
Examiner: Annika Olsson
Abstract

This study investigates the zero-waste concept applied to the food industry, and how
the environmental impact differs between a package-free scenario found at a
specialised Danish retail-store and a conventional scenario, using two reference
products typically found in normal retail stores in Sweden. In the Danish store, the
oats are sold in gravity dispensers, and the rapeseed oil is sold in glass bottles
intended to be brought back to the store by the consumer to be cleaned and reused.
Based on previous literature – both in regards to the zero-waste concept and about
packaging logistics within the food sector – a zero-waste framework is presented
intended to be used by the practitioners. The purpose of the framework is to give
general guidance to anyone aiming to offer products according to the zero-waste
principles, making sure that the products and their distribution are designed in a way
that ensures zero-waste across every parameter that exists. This shall hopefully
mitigate the risk that stores employ a zero-waste model but generates more waste,
as they for instance reduce packaging material but in turn increase food waste
generation or creates very long and inefficient transportation.
Then, a life cycle assessment of rolled oats and rapeseed oil is performed, evaluating
the package-free model found at the Danish retail store to the conventional scenario
found in Sweden. The results for oats show that even if package-free stores offer
disposable paper bags in the store – much like the system with plastic bags at the
vegetable department in most stores – the savings across most impact categories are
still large. The margin for how much food waste increase would be tolerated given
the savings is calculated, differing between about 1% for eutrophication and
acidification to 8% for climate change. For rapeseed oil, the potential savings are
not as high as for oats, and they are bound to how many times the glass bottle is
reused. If the concept were to be scaled up and the cleaning of the glass bottle
centralised, the total distance between the store and the cleaning facility should not
extend 400 kilometres, as that makes the conventional model preferable in all
categories except for freshwater eutrophication. In practise, it is advisable that the
distance does not exceed 100 kilometres to motivate the shift, as the savings
otherwise are small.

Keywords: Zero waste, package free, life cycle assessment, food waste, oats and
rapeseed oil.
Sammanfattning

Här studeras konceptet ’zero waste’, ungefär ’inget avfall’, applicerat på

livsmedelsindustrin, och hur den miljömässiga påverkan skiljer sig mellan ett
förpackningslöst scenario som kan ses i en nischad, dansk livsmedelsbutik, och ett
konventionellt scenario, där två typiska produkter i vanliga matbutiker används som
referenser. I den danska butiken säljs havregryn i gravitationsmatare (gravity
dispenser), och rapsolja säljs i glasflaskor ämnade att återföras till butiken av
konsumenten för att sedermera diskas och återanvändas.
Baserat på tidigare studier – både kopplat till ’zero-waste’-konceptet och till
förpackningslogistik inom livsmedel – så presenteras ett ramverk för ’zero waste’,
vilket är ämnat att användas av de som menar sig utöva det. Syftet med ramverket
är att ge generella råd och riktlinjer till den som påstår sig erbjuda produkter enligt
principerna för ’zero waste’, samt utifrån bästa förmåga försäkra sig om att
distributionen utformas på sätt som ser till att inget avfall uppstår i något led av
varken produktion, distribution eller avfallshantering. Detta ska förhoppningsvis
minska risken för att ’zero-waste’-butiker anammar ett tänk som i slutändan leder
till att mer avfall genereras, som att de till exempel reducerar mängden
förpackningsmaterial men vilket i sin tur leder till att mängden matavfall ökar.
Vidare så genomförs en livscykelanalys av havregryn och kallpressad rapsolja, där
den förpackningslösa modellen som återfinns hos den danska livsmedelsbutiken
jämförs med det konventionella scenariot i Sverige. Resultaten för havregryn visar
att även om butiken erbjuder engångsförpackningar – i likhet med de plastpåsar som
erbjuds i grönsaksavdelningen i de flesta butiker – så minskas miljöpåverkan stort i
samtliga påverkanskategorier. Marginalen för hur mycket mer matsvinn som skulle
kunna tillåtas utan att den konventionella modellen blir att föredra är olika stor. För
övergödning och försurning är marginalen ungefär 1%, medan den för global
uppvärmning är 8%. För rapsolja så är de potentiella besparingarna inte lika höga
som för havregryn, och de är tätt kopplade till hur många gånger glasflaskan
återanvänds. Skulle konceptet genomföras i större skala och diskanläggningen för
glasflaskor centraliseras, så får inte det totala avståndet mellan butiken och
anläggningen överstiga 400 kilometer, eftersom detta leder till att den
konventionella modellen är att föredra för alla påverkanskategorier utom
sötvattensövergödning. I praktiken är det önskvärt att avståndet inte överstiger 100
kilometer för att motivera ett skifte.
Nyckelord: Zero waste, förpackningslös distribution, livscykelanalys, matsvinn,
havregryn och kallpressad rapsolja.
Acknowledgments

After being involved in several different projects concerning topics within the food
sector, it has been very rewarding to use my engineering degree and investigate such
a current topic as the package free distribution of food.
Firstly, I want to thank Løs Market, Gunnarshögs Gård and Saltå Kvarn for their
invaluable contribution to this study. Without their help, this thesis would not have
been possible.
Secondly, I would like to thank RISE Agrifood and Bioscience (previously SP) for
their input on life cycle assessments, both in terms of knowledge and resources.
I hope this thesis can bring more clarity to the current debate about package free
distribution of food and to help retail stores to further decrease their environmental
impact.

Lund, February 2017

Vilhelm Sjölund
Table of contents

1 Introduction 10
1.1 Background 10
1.2 Research purpose 11
Purpose of the study 11
Research questions 11
1.3 Document structure 11
2 Methodology 13
2.1 Research strategy 13
2.2 Literature review 14
2.3 Survey 15
Designing a good survey 15
Identifying key aspects to include 15
Finding and identifying the stores 16
Collecting responses 17
Interpretation 17
2.4 Life Cycle Assessment 17
Goal definition 18
Project scope 19
Impact assessment 23
Completeness check and sensitivity analysis 24
Constructing the theoretical scenarios 25
Methodological considerations for life cycle assessment 25
Excluding food waste 26
2.5 Combining methods 26
3 Theory 27
 3.1 Zero waste 27
What does zero waste mean? 27
Opposing theories of sustainability 29
Zero waste and the distribution of food 30
3.2 Packaging and food 30
3.3 Perspectives on waste reduction 32
Quantifying packaging reduction benefits 32
Overview of waste generation within food distribution 33
Reverse logistics as a waste prevention measure 37
3.4 Retail stores without primary packaging 39
4 Life cycle inventory 40
4.1 System boundaries 40
4.2 The rolled oats system model 42
The conventional model 42
The package-free model 42
4.3 The rapeseed oil system model 43
The conventional model 43
The package-free model 43
5 Results 45
5.1 Packaging consumption per functional unit 45
Packaging material consumption for oats 45
Packaging material consumption for rapeseed oil 46
5.2 Life cycle impact assessment 47
Impact results 47
Impacts per packaging activity 47
5.3 Theoretical scenarios 49
5.4 Completeness check 54
5.5 Toxicity analysis 55
5.6 Sensitivity analysis 55
Conventional oat system 55
 Package-free oat system 56
Conventional rapeseed oil system 56
Package-free rapeseed oil system 56
5.7 Impact compared to the food production 57
The percental impact of packaging 57
Food waste generation 59
Possible areas for differences in food waste generation 59
5.8 Selected survey results 60
5.9 Analysis of survey results 61
Store operation procedure 61
Values and ideology 62
6 Discussion 63
6.1 Previous work 63
6.2 A zero-waste retail store framework 64
The framework 65
The product 66
The distribution 67
The packaging 68
Package free or zero-waste 69
6.3 Combining models 69
Impact results 69
Pragmatic evaluation of relative packaging impact 70
7 Conclusions 72
7.1 Research question 1 72
7.2 Research question 2 73
7.3 Research question 3 74
8 References 76
Appendix A 81
Appendix B 82
Appendix C 83
 1 Introduction

1.1 Background
During the last decade, retail stores marketing themselves as ‘zero-waste’ or
‘unpackaged’ have emerged at an increasing pace. This has led to a debate as to
whether this is the environmentally preferred direction for the food industry. Still
today, close to the end of this study, it is still a current topic in media with experts
commentating the role of packaging in food distribution (Pettersen & Hægermark,
2016; The Economist, 2016). However, few studies have evaluated the performance
of these package-free distribution models. Two noteworthy examples include Nessi,
Dolci, Rigamonti & Grosso (2016), where the authors evaluated the model for rice,
pasta and breakfast cereals, with varying results but mostly in favour of package-
free distribution, and Cleary (2013), who researched refillable wine bottles with a
centralised solution, stating that the transport distance to the washing facility is the
critical success factor.
Distributing food requires packaging, be it disposable or reusable. Even stores
aiming to minimise packaging will be bound to use it to some extent, leaving us
with different possible scenarios. Larger containers – disposable or reusable – can
be bought and used to refill smaller, reusable consumer packaging. If retail stores
use reusable packaging even at the supplier stage a reverse supply chain takes place,
adding a layer of complexity as they now also must manage outbound logistics of
the packaging containers. Because of this variety of packaging supply chain options,
it is difficult to give a definite answer as to what package free distribution of food
entails. Substituting primary packaging for reusable options is what many stores
market, but for secondary and tertiary packaging such as corrugated boxes, shrink
wrap and wooden pallets, less knowledge exists as to how the stores operate.
A central question of package free distribution is whether it is justified when food
waste is considered, as packaging can serve to protect the food product and therefore
reduce waste generation. No previous studies have evaluated a package free food
distribution and at the same time quantifying the effects on food waste generation.
To answer the question which packaging model is preferable in terms of
environmental performance, one must both calculate the impacts of the packaging
itself and differences in performance. But obtaining robust data is difficult and
collecting a few samples would not lead to satisfying results.
10
 1.2 Research purpose

Purpose of the study

The primary purpose of this study is to further the understanding of the potential
benefits or drawbacks that could come from redesigning the supply chain model in
the food sector towards removing packaging in favour of reusable, ‘package-free’
options. The study shall also provide a comprehensive review of the emerging
concept where stores offer customers the possibility to buy food without packaging.
The concept shall be described in its current state, compared to research within the
area, evaluated based on environmental performance and produce a framework that
can be used by the stores which both could unify the concept and increase the
potential environmental benefits from their operations.

Research questions

The following research questions shall be answered:

RQ1: Which definitions of ‘zero waste’ currently exist within food
distribution; are they contradicting each other, and can a general definition
be found which can be used as a framework for the practitioners?
RQ2: For rolled oats and cold pressed rapeseed oil, what is the difference
in environmental impact between the ‘package-free’ distribution model and
the conventional, packaged distribution?
RQ3: In terms of environmental performance, which positive and negative
aspects exist when grocery stores adopt a package-free model and in which
ways could a store benefit by combining the different models?

1.3 Document structure

First, the methodology of this thesis is presented. Then, in chapter three, the results
from the literature review used to answer both the first and third research questions
is presented. In chapter four, the inventory data from the studied supply chain
systems are illustrated, and the impact assessment results are presented in chapter
five.
Chapter seven is dedicated to provide an answer to the first and third research
question. To answer the third, results from both the life cycle assessments, the

11
survey and the literature review are used, and therefore all discussion relating to this
can be found in this chapter.
Finally, the main conclusions are presented along with recommendations for further
research.

12
 2 Methodology

2.1 Research strategy

When assessing the environmental impact of a system, the life cycle approach is the
preferred methodology, standardized by ISO (ISO, 14040:2006) and with
methodological recommendations from the European Commission (European
Commission , 2010). The life cycle methodology is closely related to that of a case
study, both ensuring quality through validity, reliability and with sound
generalization principles (Yin, 2014; European Commission , 2010). Many studies
apply both terminologies, either describing their research as a case study with a life
cycle approach (Paolotti, Boggia, Castellini, Rocchi, & Rosati, 2016), as a case
study including a multi-method approach which includes life cycle assessment
(Väisänen, et al., 2016) or as a general life cycle study with a regional case study
approach (Yay & Suna, 2015). However, common to these studies is that the
methodology is a life cycle approach and not affiliated with case study research.
This study will therefore build upon this convention, applying the life cycle
approach with a case study of the zero-waste model in a Danish and Swedish
context.
A summary of the chosen research strategy is illustrated in Figure 1. The first
research question will be answered by reviewing literature combined with a survey
to the stores. The second research question will be answered through four
independent life cycle assessments. The third research question will be answered by
combining the results from the two previous studies, combined with additional
literature findings. Results from the life cycle assessments are tied to a single store,
as they do not address several zero-waste stores and how they operate. By
combining these results with findings from the survey, more general conclusions
can be drawn about the zero-waste model at large and which effects on
environmental performance can be expected if a store chooses to deliver products
accordingly.

13
Figure 1: Summary of research methodology

2.2 Literature review

In Figure 2, the initial search strategy is summarised. The purpose of this literature
review was to get an initial understanding of the field and create a theoretical frame
of reference. In the beginning, it was unclear which keywords were relevant as no
clear definition of the concept exists. To overcome this issue, a citation pearl
growing was conducted (Rowley & Slack, 2004), involving an unstructured search
for a few key articles and identifying relevant keywords for further search from
these.
For the first research question, a comprehensive review of previous zero waste
research was found during the first literature search process. The articles listed were
evaluated based on their appropriateness to food packaging logistics, presented as
inclusion criteria in Figure 2.
However, over the course of this study, literature was searched for continuously and
collected from various sources. Several articles have been provided by the
supervisors, combined with input from other person from academia and the industry.
Furthermore, the bibliography from articles found during the primary search in
Figure 2 have been examined to improve the theoretical framework.

14
 Source Inclusion criteria: either or
Relates to zero waste frameworks within
retail or with generalizable applicability
RQ1
Relates to economic theories
Databases LUB Search with a focus on sustainability
Relates to unpackaged distribution of food
RQ3 Relates to the conflict between food
and packaging waste
RQ Keywords Results After exclusion
RQ1 Zero waste & waste management 120 8
RQ1 Circular economy & packaging 47 2
RQ3 Waste prevention & packaging 256 16
RQ3 Food waste & retail & packaging 30 10
Note: Several other keywords were used, including ‘Life cycle assessment’ and ‘environmental
impact’. These searches did not generate articles other than those listed above, and has therefore
been excluded from the summary.
Figure 2: Overview of the search strategy for the literature review

2.3 Survey

Designing a good survey

A survey describes a population, either only by describing the population itself or

by comparing it to some reference group (Sapsford, 2007). To ensure good quality,
high participation, thorough reflection about chosen questions and generating
answers that gives unambiguous results are vital parts of a successful survey
(Sapsford, 2007). Since relatively little is known about the zero waste stores, the
survey was sent out after both going through the literature and visiting the store to
find out about relevant aspects to include. In addition, a qualitative interview with a
researcher was done before sending out the surveys.

Identifying key aspects to include

To find out which questions that should be included, the literature review for the
first research question was completed before working with the survey. Thereafter,
an open-ended interview was conducted with Mattias Eriksson, who is a postdoc
15
researcher at the institution for energy and technology at Sveriges
Lantbruksuniversitet in Uppsala. The purpose of these steps was to gain a better
understanding for which environmental pitfalls may exist if a store adopts the zero-
waste model. The interview largely consisted of discussions around packaging and
food waste, and whether the zero-waste stores would generate more food waste than
a regular store. This resulted in the addition of several questions relating to
packaging, such as if the zero-waste stores would consider using bio-degradable
packaging if this improves the shelf life of a product, and whether the stores use
reusable packaging further down the supply chain or just for the primary, consumer
packaging.
The questions included in the survey were either value-based or related to how the
store operated. Valued-based questions were included to see how similar the stores’
concept is to the principle of zero-waste management as explained in chapter 3.1.1.
Typical questions were if the store gave special prioritisation to locally produced
food or if they only sell organic produce. The other questions related to how the
stores operate. This information was collected to see if the Danish store that was the
basis for this study represents how the other stores operate. The full results of the
survey can be found in Appendix B.

Finding and identifying the stores

Locating the relevant stores for this study both included finding information about
the store itself as well as ensuring that it could be characterized as a store within the
description of system I, see chapter 2.4.2.4. A website called Bepakt have assembled
a thorough list of package free stores around the world which was used as a base for
this study (Muller & van Engeland, 2015). However, after evaluating the stores
listed at this site, not all were in line with the scope of this study. Furthermore,
several stores were found to be missing from this list based upon the author’s
previous knowledge, and therefore an independent search was conducted as a
complement to earlier findings.

Keywords Languages
Zero waste store Bulk store
Swedish, Danish, Norwegian, Icelandic,
Package free store Unpackaged store Finnish, English, Irish, Dutch, French,
German, Polish, Russian, Italian, Spanish,
Wholefoods store Bio-based store Portuguese, Korean, Japanese, Chinese.

Table 1: Search strategy for localising relevant stores

The evaluation process included visiting the stores’ websites and social media
platforms, looking for key markers. These were primarily images of food sold

16
without packaging and information about their vision and whether they claim to be
a store selling unpackaged food. In total, 52 stores were found. To further increase
the accuracy, a control question was added to the survey: “My store aims to remove
disposable packaging as far as possible, offering customers bulk food and the
possibility to bring their own, reusable packaging.”, with possible answers being
“yes”, “no” and “we offer mostly unpackaged food but also sell certain food items
packaged”.

Collecting responses

The survey was written using Google Forms, and initially delivered through email
to each of the stores. This resulted in a very low response rate, and several iterations
were performed, contacting stores in different ways. This included contacting them
through social media, texting and/or calling the store, and translating the messages
to their respective language in case English was too foreign to them. Despite the
efforts, only 16 stores filled out the survey. Out of these, 14 are in Europe and 2 in
North America. Obtaining statistically significant results for a population of 52
requires a large degree of participation, which has not been obtained here.

Interpretation

As the survey received too few responses to be statistically significant, these results
will not be used to validate theories about the zero-waste concept at large, as those
conclusions would rely on very weak premises. This means that the results from this
life cycle assessment cannot be generalised well to all other stores as they might
operate in different ways, which would render different results. The results can
however give some insight about the concept itself and how stores operate. This will
be used in conjunction with results from previous studies to better understand the
concept.

2.4 Life Cycle Assessment

The methodology for conducting life cycle assessments has been standardized by
ISO, and the general approach is shown in Figure 3. The ILCD handbook (European
Commission , 2010) offers a thorough methodological guidebook for conducting
life cycle assessments. The research strategy for this study is largely based on this
handbook, complemented with a few other sources of literature (Baumann &
Tillman, 2004; Rydh, Lindahl, & Tingström, 2002).

17
 Figure 3: Overview of a life cycle assessment (ISO, 14040:2006)
In recent years, more emphasis has been put on the assessment as an iterative
approach (European Commission , 2010). The general idea is to collect initial data
and construct the life cycle model at an early stage. Thereafter, an analysis is
performed to assess if the scope definition should be revised or if more accurate data
from foreground processes are needed. Then another iteration is performed,
validating the new data, and building an improved model. This study will be limited
to one iteration, as both time and money for field visits are limited.

Goal definition

2.4.1.1 Intended applications

This study intends to perform four individual life cycle assessments, evaluating the
environmental impact associated with two different supply chain systems for food
distribution, namely:
1. System I, hereby denoted the ‘package-free model’, characterized by a far-
reaching ambition to reduce waste by eliminating disposable packaging in
favour of reusable packaging solutions. A niched type of small to medium
sized retail store employ this model.
2. System II, hereby denoted the ‘conventional model’, characterized by a
vision to not use more packaging than necessary while still maintaining the
functions of packaging, e.g. to protect the product, provide information to
the customer and market the product. This system comprises most western
retail stores.
The purpose of the study is to evaluate how the environmental impact differ for the
two different supply chain models. To determine superiority in terms of
environmental impact, an externally reviewed comparative assertion is the
recommended approach (European Commission , 2010). A comparative assertion
intended to be disclosed to the public entails several mandatory requirements in
terms of execution, documentation and reviewing (ISO, 14040:2006; ISO,
18
14044:2006). This study does not intend to perform an analysis which can be used
as a foundation for claims of superiority between the two supply chain models, and
the requirements dictated by ISO are not met.
The four individual assessments shall calculate the environmental impact of each
system as is, meaning that no corrections will be made to adjust for different
distances to packaging suppliers or in any way make the systems better resemble
each other. To evaluate the impacts adjusted for differing variables, different
theoretical scenarios will be computed in connection to the sensitivity analysis.

2.4.1.2 Limitations
Package-free stores target a specific customer segment and have a relatively small
range of products, whereas conventional stores target a broad segment of customers
and usually offer a wide range of products. The conclusions drawn based upon a
niche market generally have poor generalisability to an industry at large, which is
accounted for in this study (European Commission , 2010), and conclusions are
generally limited to encompass other package-free stores. Despite these limitations,
the question about which supply chain model is environmentally preferable still
exists, and this study will provide valuable input to which system is better. The
limited applicability mostly concerns that the concept may not appeal to a larger
crowd, that customers may not bring their own packaging to the store and that
customers may not fill their packaging properly, leading to increased food waste. In
case the package-free model is preferable, overcoming these obstacles would make
the results generalizable to a broader market.

2.4.1.3 Decision context and target audience

This study will create a micro level decision support for each of the two studied
distribution models for the studied food products. It will determine the impact for
each product flow with its corresponding processes and activities.
This decision support is intended to be used by the research field of packaging
technology and the concerned industry parties: retail stores, primary food producers
and packaging companies.

Project scope

2.4.2.1 Functional unit

Two different functional units are studied:
1. The packaging distribution of one kilo ready-to-eat rolled oats to the end
consumer, including end-of-life treatment.
2. The packaging distribution of one litre ready-to-eat rapeseed oil to the end
consumer, including end-of-life treatment.
19
The functional unit does not consist of the food product itself or the distribution of
it, but the production and distribution of any packaging material that relates to the
distribution of the food product itself, and the subsequent end-of-life management
the packaging material. Any food waste generation is not accounted for, as no
reliable data could be collected. For this discussion, see chapter 5.7.2.

2.4.2.2 Deliverables
For each of the studied food products, a life cycle assessment study will be
delivered, including impact assessment and interpretation. Addition to this,
theoretical scenarios, which seeks to answer under which circumstances each of the
studied systems perform better than the other, shall be delivered.

2.4.2.3 Life cycle inventory (LCI) modelling framework

For all four of the studied systems, attributional modelling will be used as the
inventory modelling framework. This means that all environmental inputs and
outputs are summed up for processes that spans from raw material extraction to end-
of-life treatment, or in easier terminology, attributes the sum of environmental
burdens to a specific functional unit within a certain system boundary. The other
common system model is called change-oriented or consequential LCA, which
instead looks at which impact a change in a system would cause.
A central question in attributional modelling is how to handle cases of
multifunctionality, for instance when the production of a product also generates
several co-products. For the main processes, no multifunctionality exists as the
functional units only provide a single function, i.e. to be delivered to the end
consumer. However, within the processes themselves, cases of multifunctionality
exist, where the actual production of a product generally leads to the production of
several co-products. An example is the production of sodium hydroxide that stems
from a chlor-alkali electrolysis process which also outputs chlorine. To handle this,
Ecoinvent offers two different system models: allocation cut-off by classification
and allocation at the point of substitution (Ecoinvent, 2014). For all assessments and
processes in this report, the cut-off model has been used. In this model, recycled
content is available burden-free for primary producers, or in other words, the
recycled material only carries the impact of the recycling processes themselves. For
most activities used, by-products are cut-off from their producing activities.
Other than multi-output activities, the waste management of packaging materials
consists of several cases of multifunctionality, either in the form of electricity
generation through municipal waste incineration, or raw material reuse through
recycling. To correctly model recycling and to allocate avoided burden, one must
consider recycling rates, which level of material downgrading occurs and the
inventory and impact of the subsequent processes that relates to sorting,
transporting, and processing the collected materials until they have achieved an

20
equal quality to virgin materials (Baumann & Tillman, 2004; European Commission
, 2010). For attributional modelling with net positive end-of-life products, the
recommended approach is to determine an average inventory load based on infinite
cycles of waste management. Each cycle will eventually converge to zero unless the
system has both a perfect collection and zero downgrading, which in practise never
occurs (European Commission , 2010). However, many studies adopt a simplified
approach including only the first iteration (Gaudreault, 2012), which is adopted here
as well. This choice was motivated based on the methodology proposed by ILCD,
shown in equation 1, where e annotates the average inventory per unit of material,
p the primary amount of material, I the amassed LCI impact for the total amount of
materials based on the recycling grade, and U the total mass of all recycled content.
P, W, and R annotates primary, waste and recycling LCI efforts. For instance, if we
have one kilo of primary material (p), with a recycling grade of 90 %, the total mass
would be 10 kg of material.
𝐼 𝑝 ∗ (𝑃 + 𝑊 + 𝑅 ∗ (𝑟 − 𝑟 𝑛+1 )
𝑒= = (𝑒𝑞1)
𝑈 (1 − 𝑟) ∗ ∑𝑛𝑖=0 𝑝 ∗ 𝑟 𝑖
However, when simplifying the expression, we end up with equation 2 (see Europan
Commission (2010), annex C for a detailed description of the mathematic
approach).
𝑒 = (𝑃 + 𝑊) ∗ (1 − 𝑟) + 𝑅 ∗ 𝑟 (𝑒𝑞2)
Modelling this would often be meaningful, but as R is defined as “LCI of effort for
reuse/recycling/recovery per unit of material, part or energy carrier” (European
Commission , 2010), W would be negligible for packaging processes considered
since plastics and paper eventually end up in waste incineration plants (and therefore
belong to R), and glass and steel have negligible leaching effects when sent to inert
material landfill (Levova, 2016; Hischier, 2016). The impacts from the average
inventory does not differ from a single recycle chain, since no further waste occurs,
other than what is recycled in the first place.
Recycled material was assumed to replace the virgin production of materials in this
study. Wherever the Ecoinvent cut-off model use recycled content as burden-free
inputs, these activities have not been awarded recycling credits, as that would lead
to double-counting. Another issue is when packaging is produced in one country
and then exported to another country where it is recycled. Here, the numbers for the
actual secondary material use in the production factory were used, meaning that no
additional credits were awarded if the recycling rates in the country where the
recycling takes place are higher. In practise, only polyethylene shrink film was
awarded material credits.
In conclusion, a closed-loop recycling model - addressing downgrading and
incomplete recycling rates - where materials are assumed to substitute the same
production line has been assumed.
21
 2.4.2.4 System boundaries
When using secondary data, it is important to control which upstream processes are
included in the targeted process. For instance, the production of LDPE pellets –
which can further be processes into stretch film – includes not only the actual
production of the pellets but also the extrusion of crude oil, e.g. a cradle to gate
perspective for this product (Plastics Europe, 2008). If the practitioner is not careful,
this may lead to double-counting of processes, invalidating the results of the study.
The Ecoinvent data provides a detailed description of what is included in a unit
process to prevent double-counting on the form "From reception of raw materials
[possibly further specified] at the factory gate [possibly further specified]" or
"Service starting with the input of [e.g. labour and energy]” (Weidema, et al.,
2013), making it easier to avoid this error.
The ILCD handbook strongly recommends a quantitative cut-off criterion
(European Commission , 2010). This sounds good in theory, but presents several
practical issues. The methodology is difficult to implement in its current form,
because when initiating the assessment, no knowledge exists about the expected
level of completeness or even which processes needs to be included to achieve this
level. This can be overcome by adopting an iterative approach to the assessment,
updating the judgement of completeness along the course of the assessment.
However, a more severe issue is that the Ecoinvent database does not follow any
strict quantitative cut-off rule (Weidema, et al., 2013). The datasets are as complete
as the knowledge of the data providers allow. This means that even though a
quantitative cut-off rule were to be adopted in this assessment, the uncertainty of the
impacts would still be bound to the uncertainties found in the Ecoinvent dataset. To
overcome this problem, the datasets of the two supply chain models will include as
similar entries as possible for the different processes to avoid completeness
irregularities between the models.
Initially, a cut-off criterion was intended to be used where sub-suppliers of
packaging material would be cut off when the contribution did not exceed five
percent. But over the course of the study, system expansion was used iteratively to
enhance the accuracy of the model, rendering the cut-off unnecessary as the supply
chain systems were modelled from cradle, or in other words that all packaging sub-
suppliers were modelled. For remaining processes, a cut-off at one percent was used,
and processes were evaluated to see if they could lead to a higher contribution. An
example of processes cut off is the electricity consumed by the forklift in the
warehouse. Even though a process might contribute with a small share of an impact
factor, it might have a large contribution in some other aspect. An example is highly

22
toxic materials that may be overlooked if a quantitative cut-off criterion is used. For
the excluded processes, no risk for this error could be identified.

Impact assessment

The relevant midpoint impact assessment categories have been chosen based upon
the recommendations set out by the European Platform on Life Cycle Assessment
(Europan Commision, 2011). In Ecoinvent, the database layer computes the values
for the different assessment categories, and the ILCD categories were added to the
third version of the database (Hischier, et al., 2010).

Quantitative Qualitative Not

Assessment category Unit
analysis analysis included

Climate change Kg CO2-Eq X

Ozone depletion Kg CFC-11 X
eq/FU
Human toxicity X
Particulate matter Kg PM2.5 X
eq/FU
Ionizing radiation – damage to human kg U235-Eq X
health
Ionizing radiation – damage to ecosystems X

Photochemical ozone formation Kg C2H4-eq X

Acidification Mol H+-Eq X

Terrestrial eutrophication Mol N-Eq X

Freshwater eutrophication Kg P-Eq X

Marine eutrophication Kg N-Eq X

Land use X

Depletion of abiotic resources X

Table 2: List of the chosen impact assessment categories

The climate change category is based on the global warming potentials developed
by the Intergovernmental Panel on Climate Change (IPCC). They calculate these
based on three different time frames: a 20, 100 or 500-year cut-off period (Myhre,
et al., 2013). Even though the 500-year period is the most robust and recommended
praxis, the 100-year cut-off period is widely accepted and used as a basis for the
Kyoto Protocol I . There is no scientific argument choosing a 100-year cut-off period
over other periods, and IPCC stresses that this choice is a value based judgement
(Myhre, et al., 2013). In this report, the 100-year cut-off period will be used since it
is also used by previous research and therefore makes this report more comparable
to these reports.
23
For the category ionizing radiation damage to ecosystems and depletion of abiotic
resources, ILCD deems the underlying methodology to be interim, and it has
therefore been excluded from this study. (Europan Commision, 2011; Garnier-
Laplace, et al., 2006).

Figure 4: The difference between sensitive and non-sensitive areas as modelled in the AE
method (Margni, et al., 2008).
The chosen methodology for measuring acidification is one called accumulated
exceedance (Margni, et al., 2008). This method provides country-dependent
characterization factors, which gives an average value for the buffer capabilities of
a country. For Swedish conditions, the buffer conditions vary greatly. For instance,
the limestone environment of Gotland has far greater buffer capabilities than the
acid moors of Jämtland, but the methodology is still the most accurate to this date
(Europan Commision, 2011).

Completeness check and sensitivity analysis

During the interpretation phase, a completeness check and a sensitivity analysis will
be done.

Figure 5: Sensitivity analysis matrix

(European Commission , 2010)

24
The sensitivity analysis will be undertaken using the sensitivity matrix shown in
Figure 5. The focus will be on high priority items where the underlying inventoried
data has low quality and the value of the process largely affects the outcome of the
study.

Constructing the theoretical scenarios

For both system models, theoretical scenarios were computed based on observations
from the Danish retail store. In general, transport distances and emission classes of
vehicles were normalised to not give benefit to either system. Firstly, car transports
to the consumer were excluded as they contributed to a major share of all impacts,
and do not relate to the difference between each system model. Secondly, electricity
credits from waste incineration were excluded due to the different electricity
composition between Denmark and Sweden. Including them greatly skewed results
for carbon emissions in favour of Swedish management due to fewer coal plants,
and ionizing radiation in favour to Denmark due to the nuclear power plants in
Sweden.
The best-case scenarios were chosen based on a discussion with the owner of the
Danish retail store owner. The poorer scenarios, for instance assuming fewer usage
cycles of the glass bottle in the rapeseed oil system, were chosen by the author to
shed light on relevant weaknesses in the package-free distribution model.

Methodological considerations for life cycle assessment

Even though it is highly recommended to include every impact assessment category

available (European Commission , 2010), it is common that assessments performed
on food products address a subset of these categories (Sonesson, Cederberg, Flysjö,
& Carlsson, 2008; Wallman & Sonesson, 2010; Cederberg & Nilsson, 2004). This
choice can be well motivated if the studied system is not affiliated with certain types
of emissions (European Commission , 2010), but the relevant impact categories for
a food product may not be the most relevant for the packaging itself (Silvenius, et
al., 2014). It’s unclear how to evaluate the impacts between food products and
packaging materials when they result in emissions for different impact categories,
as no weighting between different impact categories could be identified (Europan
Commision, 2011; Baumann & Tillman, 2004; Rydh, Lindahl, & Tingström, 2002).
When looking at waste prevention initiatives of packaging, ecotoxicity, metal
depletion, freshwater eutrophication and ozone depletion are the categories
experiencing the most significant improvements (Cleary, 2014). For the production
of food, it is unlikely that for instance metal depletion would contribute
significantly, and evaluating which system would be preferable is then difficult as

25
they perform well in different categories. The issue remains unresolved in this report
as well.

Excluding food waste

To give a definite answer to which distribution model is preferable, one must

consider differences in food waste generation. As will be discussed in chapter 3.3.2,
previous studies have not quantified food waste when comparing different
packaging solutions. For this study, it was decided not to include food waste as
obtaining high enough quality of data was deemed too difficult given the project’s
time frame and the required efforts to complete the other parts. To overcome this
limitation and still contribute to the academic research, the impact of the packaging
will be evaluated against the impact of food production, and a tolerance for food
waste will be computed. This means that given the reduction of emissions for either
system, should it be accompanied by X percent increase in food waste, the overall
impact would be higher even though the impact from the packaging itself is lower.

2.5 Combining methods

A life cycle assessment only looking at a single zero waste store is difficult to
generalise, but employing a mixed method approach can partly overcome this
limitation. General, epistemological issues are still unresolved as the dualism
between subjectivism and objectivism collide when attempting to combine methods.
A few philosophical theories are presented to overcome this dualism, such as
adhering to the dialectic or pragmatic perspective (Greene & Hall, 2010). In this
study, a pragmatic approach seemed most appropriate and in line with engineering
conduct.
To answer the third research question, a combination of the results from the survey,
the life cycle assessments and literature findings will be used. This means that the
benefits and drawbacks of the zero-waste model when applied to conventional stores
will be discussed, and how the results from this life cycle assessment can be
interpreted given limitations to generalisability.

26
 3 Theory

This section includes the literature findings used to answer the first and third
research questions. The first section describes what zero waste means and how it is
– or could be – applied in a food logistics context. The second and third chapter
relates to the role of packaging within food logistics, describing the general purpose
of packaging, how it’s used, which unresolved issues still exist, and which end-of-
life treatment options exist. Finally, findings from a study that specifically targeted
zero-waste stores is summarised.

3.1 Zero waste

What does zero waste mean?

The zero-waste movement originates from the 1970s where it was coined by Paul
Palmer, a doctor in chemistry from Yale University (Zaman, 2015; Palmer, 2007).
It aims to minimize any waste generation, including emissions, by creating products
that can essentially be reused indefinitely. However, the author agrees that this
approach is infeasible and proposes a more pragmatic standpoint towards the life
time of products by moving away from the current throwaway, consumeristic
society (Palmer, 2007). Over the years, a variety of definitions have been used for
zero waste, and even though attempts have been made to standardize the concept
(ZWIA, 2009), many different definitions can be found. Notable examples include
complete diversion from landfill and incineration (Welsh Assembly Government,
2010), diversion from landfill but permitting incineration ( (Abbasi, Premalatha, &
Abbasi, 2012; Björk, 2012; Premalatha & Tauseef, 2013) as cited in (Zaman, 2015))
or in conjunction with the concept circular economy (Europan Commission, 2014).
A summary is given by Bartl (2014), proposing a further division of the term to
avoid this ambiguity.
By subdividing ‘zero-waste’, implementing a zero-waste strategy enables the use of
different phases depending on the current waste management situation. Different
countries face different types of waste management challenges.

27
For some, a diversion from landfill strategy could be fruitful, whereas in a
Scandinavian context where waste is incinerated or recycled rather than landfilled,
moving further down towards a zero-waste society would be more important. Zaman
(2015) proposes a three-phase implementation:
1. The first phase includes shifting industry to adopt cradle-to-cradle design
and similar resource extraction principles.
2. The second phase targets consumer behaviour, endorsing customers both to
consume consciously in terms of quantity and with regards to the impacts
of each product’s production and waste management.
3. The third phase involves a political shift towards the circular economy by
creating green job opportunities and strategic plans.
Suggested term Definition
Zero landfill Total or almost total diversion of waste from disposal (e.g. landfill); Feasible options
include other recovery (i.e. incineration) as well as recycling.
Zero incineration Total or almost total diversion of waste from disposal as well as other recovery (e.g.
incineration); All waste that is generated needs to be recycled.
Zero waste Total or almost total avoidance of waste by excessive waste prevention and re-use;
landfill, incineration as well as recycling must not take place.

Table 3: Suggestions for a precise wording in the field of zero waste (Bartl, 2013) as cited in
(Bartl, 2014)
Palmer (2007) argues that the original intention of zero waste more closely
resembles the current state of literature, but for decades the term has been used as
diversion from landfill or seen as an inapplicable model due to the unreasonableness
of a society entirely freed from waste. In previous years, the concept has regained
its meaning in line with authors criticising earlier implementation strategies that
usually interpreted zero waste as diversion from landfill. An example is Krausz
(2012), who stresses how the local authorities are affected as the responsibility for
end-of-life management generally falls on them. He argues that we need to localise
production as local communities otherwise can’t prevent hazardous waste from
entering their waste streams; de-emphasise recycling due to the ‘feel-good’ aspect
which hinders further transition towards zero waste; and ignore percentage
diversion rates, instead introducing the new unit of measurement per capita waste
to landfill, ultimately targeting 0 kg/person/year. Prioritising local production means
that companies are affected by national laws which makes it easier for them to exert
influence on companies.
Scenarios Waste per capita Diversion from Per capita waste to
landfill landfill
Country A 250 kg 10 % 25 kg
Country B 500 kg 5% 25 kg

Table 4: Illustrative example of the per capita measurement for diversion from landfill
28
This line of argument corresponds to the waste framework directive set out by the
European Union (2008), where prevention and preparation for re-use are preferred
alternatives to recycling, illustrated in Figure 9 in section 3.3.2.

Opposing theories of sustainability

While closing in on a consensus for what zero waste is, disagreement exists to
whether it is the soundest strategy for future resource management. Another
noteworthy concept is that of eco-effectiveness (Braungart, McDonough, &
Bollinger, 2007; Dyllick & Hockerst, 2002). The critique against the zero waste
concepts is that it decouples the relationship between economy and ecology,
aggravating economic growth. Eco-effectiveness instead forms a supportive
relationship with future economic growth by promoting upcycling, meaning that
materials maintain their status as resources for a longer period. The authors advocate
an increased producer responsibility for end-of-life treatment, synergistic effects
from intelligent materials pooling which means that a central resource bank
manages material flows and ensures that they are reused multiple times, and net
positive inventions that effectively compensate for emissions by cleaning the
environment upon usage. They argue that a zero-waste approach aims for a complete
minimization of waste and emissions (ZWIA, 2009), whereas an eco-effectiveness
approach seeks an optimum between economic growth and sustainability.

Note: Author’s contribution marked in green.

Figure 6: The difference between eco-effectiveness and zero waste (Ellen MacArthur
Foundation, 2012). Theory adapted from (Braungart, McDonough, & Bollinger, 2007).
29
In practice, the technology to shift towards these upcycling strategies exist, but as
few producers adopt a cradle-to-cradle perspective products are wrapped in a mix
of different packaging materials which obstructs this transition (Hopewell, Dvorak,
& Edward, 2009; Aarnio & Hämäläinen, 2008). Advocates of zero waste and eco-
effectiveness agree on the problem with sub-primary recycling and the principle of
cradle-to-cradle design principles to overcome this issue, but goes to different
lengths in reducing material input flows, illustrated in Figure 6 (Zaman, 2014;
Braungart, McDonough, & Bollinger, 2007).

Zero waste and the distribution of food

Few studies have looked at food distribution in a zero-waste context, but Jurgilevich
et al. (2016) provide an overview of the food sector in a circular economy
perspective. It mostly addresses the production of food, but states that we need to
shorten distribution distances and balance packaging in a way that balances mass
reduction, prolonged shelf-life and protecting the food for safety reasons. But
emphasizing cradle-to-cradle design or reusable products makes less sense in the
case of food distribution, as food is downgraded once having been eaten by the final
consumer (Ellen MacArthur Foundation, 2013). Still, a key aspect can be found that
relates to zero waste and a circular economy, namely the cradle-to-cradle
perspective of the food product itself (Genovese, Acquaye, Figueroa, & Lenny Koh,
2014; Halloran, Clement, Kornum, Bucatariu, & Magid, 2014). The authors
highlight that a cradle-to-cradle perspective means converting waste to resources
within the primary production of food, better distinguish between food surplus –
which is unsellable food that is still edible – from food waste, and finding more
effective end-of-life treatment options for food.

3.2 Packaging and food

‘Packaging is a means of ensuring safe and efficient delivery of the goods in sound
condition to the ultimate consumer, supplemented by efficient reuse of the packaging
or recovery and/or disposal of the packaging material at minimum cost.’ – The
purpose of packaging (Leigh, Jonson, & Smith, 2006)
Packaging serves several purposes, including the protection of the product, as seen
in the purpose listed above. But in general, three purposes can be identified: to
protect the product, to be effective in terms of economic, environmental and
convenience aspects, and to provide information, such as for marketing,
distribution, or about additional service offerings (Leigh, Jonson, & Smith, 2006).

30
The protection is especially relevant for a perishable product such as food, where
packaging can aid in reducing product waste (Lindh, 2016).

Figure 7: The three levels of packaging (Lindh, 2016)

Figure 7 illustrates the three level of packaging that exists. The first serves to
directly protect the product, and the two additional levels provide extra protection,
stability and makes handling of a single product more efficient. Packaging systems
are generally organised in this way.
Within the food industry, packaging logistics both entail the packaging itself and
the distribution of the product, with main process stages illustrated in Figure 8. The
packaging shall protect the product throughout all stages illustrated, ensure a long
durability in storage – both in the store and at the final consumer - as this aids in
mitigating product loss, and it should be convenient to use for the consumer. The
scope of the packaging purpose goes beyond aspects related to environmental
performance. The disagreement between conventional supply chain models and
zero-waste stores exists both in how much packaging is necessary to minimise food
waste, and whether packaging should be motivated from an economic perspective
such as in the case for marketing. The zero-waste concept aims to minimise any
packaging that flows through this network and de-emphasize recycling in favour of
reusable materials, as discussed in chapter 3.1.1.

Figure 8: Overview of activities linked to packaging logistics of food. Adapted from (Lindh, 2016)
with input from (Dias & Braga Junior, 2016) & (Beitzen-Heineke, Balta-Ozkan, & Reefke, 2016).

31
 3.3 Perspectives on waste reduction

Quantifying packaging reduction benefits

Environmentally Includes food waste

preferable scenario discussion

Prevention
Baseline

Neither
No Yes
Food
Source product
(Nessi, Dolci, Rigamonti, & Grosso, 2016) Pasta X X
(Nessi, Dolci, Rigamonti, & Grosso, 2016) Rice X X
(Nessi, Dolci, Rigamonti, & Grosso, 2016) Cereals X X
(Pagani, Vittuari, & Falasconi, 2015) Salad X X
(Nessi, Rigamonti, & Gross, 2012) Bottled water X X
(Cleary, 2013) Wine & spirits X X
(Toniolo, Mazzi, Niero, Zulani, & Scipioni,
2013)
Generic X X

(De Monte, Padoano, & Pozzetto, 2005) Coffee X X

(Büsser & Jungbluth, 2009) Coffee X X

Table 5: A review of previous literature within packaging waste prevention

The general results from previous waste prevention studies indicate that
environmental benefits can be achieved by reducing the amount of packaging used,
as seen in Table 5. These studies compare a baseline scenario – often the
predominant packaging solution found on the market – with various waste
prevention scenarios. The studies have different methodologies and the waste
prevention scenarios evaluate different parameters, such as increased packaging size
or reusable packaging, but comparing a baseline with a prevention scenario is
common to all studies. The studies relate to which environmental impact difference
occurs when reducing packaging materials, and have been chosen since they relate
to the case for zero-waste stores. As seen in the left column, most studies favour the
removal of packaging in line with the zero-waste principle. For pasta, the difference
is negligible. However, none of these studies account for a potential increase in food
waste generation. Büsser and Jungbluth (2009) discuss this issue in their paper, but
have not quantified the effects.
Nessi et al. (2016) study the case of gravity bin dispensers, where retail stores buy
large containers and repack it into the dispensers. Consumers thereafter buy the

32
product through low density polyethylene bags provided by the store. The
possibility to bring their own, custom packaging was not considered in this study.
As seen in Table 5, few studies account for a possible increase in food waste
generation. In many cases, a better use of packaging can reduce the amount fo food
waste that is generated. A few authors have found contrary results as well, including
Gustavsson and Stage (2011), whose results indicate that packaging generally does
not reduce food waste for vegetables in the retail store, still saying that the savings
could occur at the household stage. This conclusion is investigated more closely by
Wikström, Williams, Verghese & Clune (2014) and Silvenius et al. (2014), and they
conclude that the effect of reduced food waste in the household stage must be
considered when evaluating the entire environmental impact of a packaging choice,
since different packaging choices can greatly affect the outcome of the amount of
waste generated.
But for many products, a clear trade-off between increased packaging and reduced
food waste exists which relates to the protective properties of packaging (Verghese,
Lewis, Lockrey, & Williams, 2015; Quested, Ingle, & Parry, 2013; Halloran,
Clement, Kornum, Bucatariu, & Magid, 2014). To determine the difference between
a baseline and a prevention scenario for these products, it’s imperative to consider
how much food waste is generated. Even though several studies address food waste
(Eriksson, Strid, & Hansson, 2012; Gustavsson & Stage, 2011; Quested, Ingle, &
Parry, 2013), no studies have been found where different packaging solutions are
evaluated against quantified food waste data.

Overview of waste generation within food distribution

Waste generation in general has is a prioritised issue for the European Union
(Europan Union, 2008). Their solution is for companies to comply to the waste
hierarchy, illustrated in Figure 9: The EU waste hierarchy Figure 9.

33
 Figure 9: The EU waste hierarchy (Europan Union, 2008)

3.3.2.1 The distribution stage

Food and packaging waste occurs frequently throughout the entire distribution of a
product’s life cycle. As mentioned earlier, packaging serves to protect the product,
often resulting in a food waste reduction. For many products, packaging accounts
for a small share of the total environmental impact which makes this increase in
packaging motivated (Büsser & Jungbluth, 2009; Roy, et al., 2009; Wikström,
Williams, Verghese, & Clune, 2014). But despite protective properties, the very
same packaging can at the same time lead to adverse results. If several units of food
– such as vegetables – are packed together, it is usually enough that a single item is
spoiled for the entire packaging to be discarded (Lebersorger & Schneider, 2014).
Some stores try to overcome this issue by unpacking the product and selling the
remaining items loose, but many stores still discard it whole. A similar problem
occurs in the distribution stage, where packaging can serve two purposes:
prolonging the lifetime and protect the product from physical breakage. A good
example where this is the case is for tomatoes, where both a refrigerated transport
and proper packaging can ensure a prolonged lifetime and minimal spoilage
(Eriksson, 2016). But for other packaged products, damage to the packaging during
distribution often renders the product unsellable, even though the product itself is
perfectly edible (Lebersorger & Schneider, 2014; Eriksson, Strid, & Hansson, 2012;
Verghese, Lewis, Lockrey, & Williams, 2015). This is partly because customers
reject the product in the store. This problem – the disconnection between the product
and the packaging – increases packaging waste further as the best-before date not
necessarily correspond to whether the food product is edible or not (Lebersorger &
Schneider, 2014; Eriksson, Strid, & Hansson, 2012).
The impact of transport relates to which mode of transport is used, and especially
which mode of transport is used during the last step between the retail store and
consumer, as this transport is allocated to far fewer products than in mass
34
distribution. While locally produced food can lead to lower emissions (Pretty, 2005),
small-scale distribution systems such as where customers travel to farm shops often
performs poorly compared to mass distribution models (Coley, Howard, & Winter,
2009). However, it is not determined that a local supply chain reduces impact, as
this relates to the food product itself and the mode of transport (Edwards-Jones, et
al., 2008).
Another issue occurs when a product is wasted to a very large extent or when the
waste is rather large in combination with the product having a large, environmental
footprint. Some exotic fruits sold in Sweden report retail waste levels above 50 %
(Eriksson, Strid, & Hansson, 2012), and for a meat department, minced beef alone
contributed to 17 % of that department’s total carbon footprint (Eriksson, 2015).
Relying on packaging inventions can in these cases present a false security, as more
efficient waste prevention measures could be taken, such as selling minced meat
frozen or exotic fruits canned (Eriksson, 2016). This would however conflict with
customer demand, as it changes the inherent nature of the food product.
Furthermore, for many products usually sold with packaging today, packaging does
not reduce waste to a large degree (Roy, et al., 2009), and the potential to reduce
waste by packaging design is argued to be more applicable to products with a high
ratio between the impact of food and the impact of packaging (Williams &
Wikström, 2011).

3.3.2.2 End-of-life treatment

Once waste occurs, the question arises as to which end-of-life (EOL) treatment is
the preferred choice. In line with the zero-waste model, a distinction between food
surplus and food waste must be made, as described in chapter 3.1.3. Avoiding
downgrading of food surplus to waste is integral to reduce impacts associated with
food waste (Eriksson, 2015). Selling unpackaged food would both mitigate
unpackaging procedures and decrease food waste sent to incineration at retail stores.
In Sweden, a master thesis by Andrée & Schütte (2010) found that approximately a
third of the studied retail stores send their food directly to incineration instead of
separating the packaging. But on the other hand, selling unpackaged foods makes it
more difficult to handle food surplus if it is going to be upgraded into new products
or donated to charity (Eriksson, 2016).

35
 Figure 10: Waste management framework, with waste &
surplus distinction (Eriksson, 2015).
If food surplus cannot be used, different EOL treatment options exist for the waste.
Vandermeersch, Alvarenga, Ragaert & Dewulf (2014) evaluates the difference
between anaerobic digestion and feed production as an EOL treatment. Their results
indicate that moving towards using food waste as a resource for feed production is
the preferred option, which is in line with the waste prevention and valorisation
framework developed by Eriksson (2015), illustrated in Figure 10. Eriksson agrees
that the vertical positioning of waste management options in the framework does
not illustrate superiority in a descending order. For some products – such as bread
– incineration is the preferred EOL treatment, outperforming both animal feed,
composting and anaerobic digestion (Eriksson, 2015). As the food sector consists of
such a wide variety of products, it is impossible to give a framework that captures
every unique trait of all products. The waste management framework gives general
guidance as to which EOL treatment options are preferred, and even though certain
products are exempt from this principle, options further up along the framework
generally leads to more favourable results. This means that even though the
environmental impact was found to be lower for waste prevention scenarios in Table
5, it still may be a sub-optimised solution if we consider the reduced impact that
comes from a better handling of surplus food. However, the waste management
options presented by Eriksson (2015) are rarely employed by retail stores. In a Dutch
study by van Sluisveld & Worrell (2013), the authors perform a case study,
evaluating different implementation options for packaging source reduction. Out of
these options, re-use was never practised in the case of food products. The second-
life surplus options should therefore be considered a theoretical possibility and not
as an option which needs to be accounted for in present studies.

36
For packaging materials, other than preventing it from occurring in the first place,
the endorsed EOL treatment option is reuse, followed by recycling (Europan
Commission, 2010). The problem with recycling packaging material is that
downgrading occurs and the recycling process itself causes emissions (Aarnio &
Hämäläinen, 2008; Hopewell, Dvorak, & Edward, 2009).
ASTM D5033 definitions Equivalent ISO 15270 (draft) Other equivalent terms
definitions
Primary recycling Mechanical recycling Closed-loop recycling
Secondary recycling Mechanical recycling Downgrading
Tertiary recycling Chemical recycling Feedstock recycling
Quaternary recycling Energy recovery Valorisation

Table 6: Different levels of plastics recycling (Hopewell, Dvorak, & Edward, 2009)

The preferred recycling is primary or closed-loop recycling, which means that no

downgrading occurs and the material can be recycled indefinite. However, as seen
in Table 6, when talking about recycling in general it does not only include this
perfect state of material conservation. The obstacles towards achieving a 100 %
primary recycling rate of packaging materials include changing consumer
behaviours to increase household recycling rate (Europan Commission, 2010),
extending producer responsibility, harmonising material choices and creating
practises that fit the waste producers’ needs (Aarnio & Hämäläinen, 2008).

Reverse logistics as a waste prevention measure

In many cases, the introduction of reusable packaging corresponds with reverse

logistics (Cleary, 2013; Dias & Braga Junior, 2016). In Sweden, fruits and
vegetables are often distributed in reusable plastic boxes which are then returned to
the packaging plant (Svenska Retursystem, 2016). Other than this, reverse logistics
within food distribution mostly consists of glass packaging for liquids such as beer,
fruit juice, honey, cider or vinegar (González-Torre, Adenso-Díaz, & Artiba, 2004).
The environmental performance of these systems relies on transport distances and
the trip return rate, making it more suitable for regional food supply (Cleary, 2013).
This relationship is seen in Table 7 as well, where single-trip packaging outperforms
reusable variants when journey distances increase. The results also strengthen
previous literature findings in terms of larger containers reducing emissions
compared to smaller packaging. However, consumers perceive that most food waste
that occurs in their household comes from food items going bad and packaging
being too large and difficult to empty (Williams, Wikström, Otterbring, Löfgren, &
Gustafsson, 2012). The results show that certain product categories are wasted to a
larger degree than others, in their case fruit, vegetables, dairy and prepared food.
37
Properly sized, easily emptiable packaging can therefore be superior to larger,
reusable containers as different products have different requirements in terms of
packaging qualities.

Packaging material

Returnable

Single-trip
Preferable
scenario
Return Journey
Source Product trips Distance
Wood: 15
Wooden
(ERM, 2008) Generic Plastic: Not stated Plastic pallet X
pallet
100
(Franklin Corrugated
Water 30 & 60 Not stated Plastic crate X
Associates, 2008) box
Corrugated
(Sustain, 2008) Generic 92 Not stated Plastic crate X
box
(Vogtländer, Fruit & Corrugated
20 & 30 0-2500 km Plastic crate X
2004) veg. box
(Capuz Rizo, Fruit & Corrugated
20 2600 km Plastic crate X
2005) veg. box

50 & 1000 1 l glass bottle <a 1,5 l PET

(Jungbluth, 2005) Drinks 50
km bottle
18,9 l jug X
Return:
(Detzel, Giegrich, 190 & 120
0,33 l: 25
Krüger, Möhler, km 0,33 0,5 & 0,7 0,5 1,5 & 2 l
& Ostermayer,
Drinks 0,5 l: 21
Single-trip: l glass bottle Xb Xb PET bottle
0,7 l: -
2004) 250 & 320
km
0,3 0,33 & 0,5
l glass bottle Xc 0,33 & 0,5 l
(PTR , 2002) Drinks 18-32 Not stated aluminium
0,5 l PET can
bottle Xc

a) This scenario performs better than the single-trip, but worse than the jug scenario. b) The authors
state no clear advantage to either system but describe several cases where the reusable glass bottle
performs better. c) Both reusable scenarios perform better than the single-trip, but no comparison as
to which reusable scenario is preferable is presented.
Table 7: Summary of results from reverse logistics studies (adapted from (WRAP, 2010)).

Achieving zero-waste in terms of not consuming any disposable packaging

throughout the distribution is reasonably largely characterised by a reverse logistics
system, and if secondary or tertiary packaging is reusable, it may result in a poorer
overall environmental performance of the system compared to a single-use model.

38
 3.4 Retail stores without primary packaging
It’s unclear when retail stores marketing themselves as package free first occurred,
but some pioneers were Unpackaged, Negozio Leggero and Effecorta (
(Unpackaged, 2016; Negozio Leggero, 2016; Effecorta, 2016) in (Muller & van
Engeland, 2015)). The concept is still novel, but has been spreading at an increasing
pace and has recently reached Sweden (Gram Malmö, 2016) and Denmark (Løs
Market, 2016).
Through an interview study with eleven European store owners, Beiten-Heineke,
Balta-Ozkan & Reefke (2016) have analysed the stores’ concept. They call the
stores ‘zero-packaging’ and define their core concepts as a focus on environmental
protection and packaging reduction. Through qualitative interviews with eleven
store owners, they provide a general description of the concept. They show that even
though it is a novel concept, it is growing through technological innovations and
concept development. For instance, one store has developed a stainless-steel
container, enabling liquid refilling without oxygen contamination. One store offers
a plastic-free bulk container which is now sold to other stores.
The stores differ greatly, with product offered ranging from 300-1500. Some offer
refrigerated meat and fish while some only offer non-refrigerated goods. Generally,
they do not handle the inbound logistics, which means that the packaging is not
always reusable from their suppliers. Generally, stores worked to minimise food
waste generation, and many either prepare food from expiring products or donate
surplus food to charity.

39
 4 Life cycle inventory

In this chapter, the life cycle inventory process is described. For a more thorough
description of data sources and the quality of the data, refer to Appendix C. In both
system models for the conventional scenario, it was assumed that a percentage of
all transports between the customer and the retail store were done by car, according
to a survey from Svensk Handel about Swedish buying patterns in retail stores
(Svensk Handel, 2009). This has not been assumed for the package-free scenario, as
customers walk to and from the store. The car transport constitutes a major part of
the total impact of the baseline scenario. Therefore, two main scenarios are
presented for the conventional scenarios, both with and without the car transport.

4.1 System boundaries

An overview of the applied system boundaries is presented in Figure 11 and Figure
12. The production of the food product itself is not included, as well as the
subsequent distribution and end-of-life management of it. What is included is any
packaging material that flows through each system model, from every sub-supplier
of packaging until it is collected and recycled or incinerated. Credits for recycling
are rewarded based on whether their respective Ecoinvent dataset already includes
recycled content. By doing this we effectively avoid any double-counting of inputs
and outputs.
The system boundaries are applied equally to all four systems, meaning that no sub-
supplier has been cut off from any system, in line with the cut-off criterion
formulated in section 2.4.2.4.

40
Figure 11: System boundaries for both system models of oats

Figure 12: System boundaries for both system models of rapeseed

41
 4.2 The rolled oats system model

The conventional model

The conventional system model for oats is based on a Swedish, medium-sized

producer of organic oats. Data that relates to the production site and the packaging
that directly relates to the product itself has been collected at the production facility
by a field visit. Through this field visit, the sub-suppliers were identified and could
be contacted via telephone. For most of the packaging material, primary data
directly from the producers have been used. In total, every sub-supplier has been
contacted, even though not all of them could provide all necessary information. The
most difficult data to collect related to transports, and in several cases no reliable
data could be retrieved about which environmental classification or how large the
lorries used were. To overcome this, a combination of theoretical calculations,
secondary data and qualified assumptions were employed.
The transport packaging is transported to the oats production facility, where the oats
are packed into a 650-gram large paper bag. This is in turn packed into a corrugated
box, wrapped in shrink film and loaded onto a pallet. The transport packaging is
recycled, and the paper bag that reaches the consumer has been assumed to be either
incinerated or recycled, per national statistics.

The package-free model

This package-free system model for oats is based on site specific data for the Danish
package-free store. Primary data about suppliers was collected at the Danish retail
store. Sub-suppliers were contacted via telephone, and as for the conventional
model, all sub-suppliers were contacted. The oats are packed into 10-kilos paper
sacks, loaded onto a pallet, and wrapped in shrink film. For some of the packaging
material, data from the conventional model was used. Once distributed to the retail
store, the oats are re-packed into gravity dispensers. The store offers paper bags,
which are ordered from a French provider. For this baseline scenario, it was assumed
that all purchases were performed using the paper bags provided by the store. As
many customers provide their own packaging this is not entirely accurate, but due
to the large variety of customer packaging, creating an accurate model was difficult
which is why this case has been moved to the theoretical scenarios. The store also
offers reusable cotton bags, but these could not be included in the analysis as the
production takes place in India, and any information that relates to the production
and subsequent transports was impossible to retrieve.

42
 4.3 The rapeseed oil system model

The conventional model

The conventional model for rapeseed oil is based on a Swedish, medium-sized

rapeseed producer. Data collection was performed at their production facility, where
most primary data about packaging took place. Sub-suppliers were contacted via
telephone. Most sub-suppliers were successfully contacted, even though not all of
them could provide all information.
After the packaging materials arrive to the production site, the rapeseed oil is packed
into 1-litre plastic bottles. These are packed in corrugated board boxes, loaded onto
pallets, and wrapped in shrink film. The transport packaging is recycled in all stages.
After being bought by the customer, the bottle is sent to incineration. For food
packaging, plastics is generally not produced from recycled content due to
legislation for food safety (Livsmedelsverket, 2016).

The package-free model

The package-free model for rapeseed oil is based on site specific data from the
Danish retail store. Primary data about packaging and producers was collected at
their store. Sub-suppliers were contacted via email and telephone. For this system,
several sub-suppliers could not be reached. The primary producer of rapeseed oil –
a Danish medium-sized producer – never responded despite several attempts of
contact. As knowledge about the packaging itself could be collected at the store, the
producer in the conventional model was used instead. Furthermore, the sub-supplier
of glass bottles could not be contacted as the wholesale agent would not disclose
their supplier. They instead provided the distance from their warehouse in Berlin to
the glass producer. This distance was cross-referenced with all German glass
producers, and one suited best to that distance. Further sub-suppliers were modelled
in the same way, e.g. the closest providers of packaging material were used.

Figure 13: Roll container ( (The Hub's

Engineering Pte Ltd, 2011; Capcon Ltd, 2014)
in (Elofsson, 2015))

43
The rapeseed oil is packed into 5-litre plastic containers, put into corrugated board
boxes, loaded onto pallets, and wrapped in shrink film. The plastic containers are
disposed of after use, or in other words are not part of the washing system for the
glass bottles. Afterwards the oil is distributed to a wholesaler where the oil is
unpacked and delivered on roll containers.
Once in the store, the rapeseed oil is tapped into glass bottles. These bottles are
produced in Germany, distributed to the store, and sold to the consumer. Consumers
can bring their own packaging as well, which is illustrated as a theoretical scenario
in Figure 17. After being used, the bottles are brought back to the store and the
customer is reimbursed for the initial glass bottle cost (a deposit system). The bottles
are washed in an industrial dishwashing machine, and reused. For the main analysis,
it was assumed that each bottle was reused 15 times before being sent off to
recycling. This value is an assumption and cannot be seen to perfectly represent the
system, which is why many different values are calculated, see Figure 17 and Figure
18.

44
 5 Results

Initially, the results from the life cycle assessment are presented in this chapter. The
tables are accompanied with shorter descriptions. A more in-depth analysis,
including sensitivity analysis and the comparison to food production and waste, is
found in chapter 5.6 and 5.7. Lastly, a few chosen results from the survey are
presented with a short description. For the full results of the survey, see Appendix
B. Throughout this chapter, the values for eutrophication (freshwater, marine, and
terrestrial) are often shown as an average between the three categories. This choice
was taken to make diagrams and tables easier to read. This approach has only been
applied to relative values, for instance comparing the impact from freshwater
eutrophication between the production of food and the packaging itself. The
different units between the three categories are therefore not in conflict as the
categories are evaluated against the same, and only the dimensionless relation is
aggregated.

5.1 Packaging consumption per functional unit

Packaging material consumption for oats

Oats – Package free Weight (kg) Relative amount of

packaging (%)
Paper (primary) 0,0205 101
Corrugated board box (secondary) 0,000246 1
Plastic film (secondary/tertiary) 0,000922 183
Oats – Conventional

Paper (primary) 0,0202 100

Corrugated board box (secondary) 0,0444 100
Plastic film (tertiary) 0,000504 100

Table 8: Quick overview of the amount of packaging material used in each system for oats. The
relative amount is illustrated in percent compared to the conventional model.
45
In Table 8, the amount of packaging material in kilograms is illustrated for both
systems. For readability, the relative use is shown in percent, showing how much or
less the package-free model uses compared to the conventional model. The package-
free model only uses corrugated board box for transporting the small paper bags to
the store, which explains the low contribution. Since they transport larger, paper
sacks, more shrink film is used to stabilise the pallet. The conventional model uses
corrugated board boxes, resulting in less film being required to maintain stability in
the lorries.

Packaging material consumption for rapeseed oil

Rapeseed oil – Package free Weight (kg) Relative amount of

packaging (%)
Glass (primary) 0,047822 -
PET (primary) 0,033199 56
Corrugated board box (secondary) 0,017774 68
Plastic film (tertiary) 0,000680 81
Rapeseed oil – Conventional

PET (primary) 0,059356 100

Corrugated board box (secondary) 0,026000 100
Plastic film (tertiary) 0,000844 100

Table 9: Quick overview of the amount of packaging material used in each system for rapeseed
oil. The relative amount is illustrated in percent compared to the conventional model.
As for the case of oats, the relative percentage of package-free distribution of
rapeseed oil is shown compared to the conventional model in Table 9. As the
package-free model uses a larger PET container, the weight per functional unit is
lower. At the same time, a larger amount of plastic film is required to stabilize glass
bottle pallets.

46
 5.2 Life cycle impact assessment

Impact results

Conventional

Conventional
Conventional

Conventional
Package free

Package free
without car

without car
Rapeseed –

Rapeseed –

Rapeseed –
Oats –

Oats –
Oats -
Unit
Impact category (per FU)
Climate change Kg CO2-eq 1,3E-01 7,1E-02 2,2E-02 4,5E-01 4,0E-01 2,9E-01

Acidification Mol H+-eq 6,3E-04 3,8E-04 1,7E-04 1,7E-03 1,5E-03 1,2E-03

Freshwater Kg P-eq 4,9E-05 3,7E-05 1,5E-05 1,2E-04 1,1E-04 8,8E-05

eutrophication
Marine Kg N-eq 1,8E-04 1,3E-04 4,2E-05 3,2E-04 2,7E-04 2,3E-04
eutrophication
Terrestrial Mol N-eq 1,5E-03 9,7E-04 3,8E-04 3,0E-03 2,5E-03 2,2E-03
eutrophication
Ionizing radiation Kg U235-eq 1,7E-02 1,3E-02 7,5E-03 8,2E-03 4,1E-03 2,8E-02

Ozone depletion Kg CFC-11 8,9E-06 9,2E-09 3,2E-09 8,9E-06 1,8E-08 2,0E-08

Photochemical Kg C2H4-eq 4,5E-04 2,6E-04 1,1E-04 1,1E-03 9,0E-04 7,4E-04

ozone formation
Particulate matter Kg PM2,5 1,1E-04 6,6E-05 4,1E-05 2,2E-04 1,8E-04 1,4E-04

Table 10: Summary of the total impact assessment for all system models
In Table 10, a summary of all life cycle assessments is presented, including both a
scenario with and without car transport to the final consumer, as this stage accounts
for a large share of the total impact of the system. The value for each impact category
is in their respective unit of measurement, which is described in Table 2.

Impacts per packaging activity

In Figure 14 and Figure 15, the impact per activity is shown. The references, such
as OC1, refer to the system boundary diagram in chapter 4.1. Eutrophication is
shown as the average between the values for freshwater, marine and terrestrial,
which is for readability purposes only. The values below do not include the car
transport between the retail store and the consumer.

47
Figure 14: Impact per packaging activity for both oats system models

Figure 15: Impact per packaging activity for both rapeseed oil system models

48
 5.3 Theoretical scenarios

Rapeseed oil –
Unit Oats – Theoretical, Theoretical
Impact category (per FU) conventional conventional
Climate change Kg CO2-eq 6,7E-02 4,0E-01

Acidification Mol H+-eq 3,6E-04 1,5E-03

Freshwater eutrophication Kg P-eq 3,6E-05 1,1E-04

Marine eutrophication Kg N-eq 1,2E-04 2,7E-04

Terrestrial eutrophication Mol N-eq 8,8E-04 2,5E-03

Ionizing radiation Kg U235-eq 1,5E-02 3,8E-02

Ozone depletion Kg CFC-11 8,6E-09 2,2E-08

Photochemical Kg C2H4-eq 2,4E-04 9,0E-04

ozone formation
Particulate matter Kg PM2,5 6,4E-05 1,8E-04

Table 11: Values for the conventional, theoretical scenarios of oats and rapeseed oil, used as a
baseline for comparison with different theoretical, package-free scenarios.
In Figure 16, the different package-free scenarios are compared to the theoretical,
baseline scenario, with difference in impact shown for each category. The first
category, PF baseline, refers to a scenario where previously mentioned categories
have been normalised, but no other adjustments have been made. For the 50 % fill
rate scenario, it is assumed that the customer only fills the paper bag half-full
compared to the capacity measured at the package-free store. The last scenario
considers the case where the customer brings a separate container, for which no
environmental burden has been awarded. This is to consider as a best-case scenario,
where a customer finds an old jar and reuses it indefinitely. This is just a theoretical

49
case and could reasonably not be replicated in a real context, but shows the
theoretical, maximum benefit of the system in its current state.

Figure 16: Theoretical scenarios compared to the theoretical, conventional scenario for oats

The rapeseed oil scenarios share several categories as both systems have the same
rapeseed oil supplier. In these cases, the transport distances have not been
normalized as they already were the same.

The different theoretical scenarios for rapeseed oil are displayed in Figure 17. As
with the oats, the PF baseline refers to a scenario where parameters have been
normalised but no further adjustments have been made. The first scenario considers
replacing the 5-litre plastic container with a 10-litre. The two following scenarios
consider different re-usage rates for the glass bottle in the store, where the zero-
waste baseline has assumed 15 cycles before recycling. The last case is to be
considered a best-case scenario, where the glass bottle is reused 50 times, and the
plastic container has been replaced with a 10-litre one.

50
Note: The original modelling of glass bottle assumes 15 usages.
Figure 17: Theoretical scenarios compared to the theoretical, conventional scenario for
rapeseed oil

In Table 12 and Figure 18, a more in-depth analysis of how the amount of usage
cycles of the glass bottle affects the overall performance of the system.
Impact category Value Relative to conventional scenario
Climate change 0,74 185%
Acidification 3,7E-03 248%
Freshwater eutrophication 2,3E-04 204%
Marine eutrophication 7,2E-04 270%
Terrestrial eutrophication 7,8E-03 303%
Ionizing radiation 5,7E-02 149%
Ozone depletion 7,1E-08 329%
Photochemical ozone formation 2,2E-03 242%
Particulate matter 3,5E-04 190%

Table 12: Theoretical worst-case scenario for the package-free model of rapeseed oil, assuming
a single usage cycle of the glass bottle before disposal.

51
 Note: The horizontal axis shows how many times the bottle is assumed to be reused.
Figure 18: Theoretical package-free impact compared to conventional, setting
glass bottle reuse as a variable.
In Table 12, we see that the worst-case scenario when the bottle is not reused at all
but only bought once and then disposed of results in about two to three times
increase in overall impact. As the glass bottle is reused, the impact per functional
unit decreases, as shown in Figure 18. For the impact categories climate change and
particulate matter, we see that the package-free model outperforms the conventional
model after three or four use cycles, whereas for acidification, photochemical ozone
formation and eutrophication, we find a threshold to be at five use cycles.

52
Figure 19: Theoretical point of break-even between conventional and package-free assuming a
centralised washing facility, equalising the models’ impact with transport distance as variable.

Figure 20: Same figure as Figure 19 but with freshwater eutrophication omitted for clearer
scaling.

53
Finally, Figure 19 and Figure 20 illustrates what would happen if we assume a
centralised washing facility in place of the in-store solution found here, and when
the emissions from transporting the glass bottle back and forth to this washing
station are equal to the emission reduction achieved by adopting the package-free
model. In our baseline scenario, we assumed 15 use cycles. This means that
depending on which impact factor we look at, the conventional model will
outperform the package-free model if the washing facility is placed further away
than between 100-500 kilometres or 1200 kilometres for freshwater eutrophication.

5.4 Completeness check

All packaging materials have been modelled from cradle to grave, and no material
flows have been missed.
For both oats systems, no reliable data could be obtained for the paper bag
production process, and it has therefore not been included in the analysis. The
impact from the paper bag is therefore slightly higher in reality than what is
presented in this study, both due to the omitted production processes and that the
amount of paper material required to create one bag most likely is higher as not all
packaging paper material is utilised during the paper bag production process.
As the process is excluded for both systems, the possible difference from different
paper utilisation between the paper bag (conventional) and the paper sack (package-
free) is assumed to be negligible.
For the package-free rapeseed oil system, primary data is missing for several
different packaging processes. Firstly, the entire production and packaging process
have been modelled based on a different rapeseed oil provider. This could affect the
outcome of the impact results for the system itself, but for the comparison between
the systems it rather enhances the results as the comparability between the systems
increases. Secondly, all upstream processes from the Berlin packaging glass
wholesaler have been approximated. It is not known if any relevant material flows
have been missed. These processes have been modelled based on the primary data
obtained from other companies throughout this study, and the comparability is
assumed to be good. Typical values estimated or calculated are the amount of shrink
film needed to wrap a pallet or how many corrugated board boxes fit on a pallet.

54
 5.5 Toxicity analysis
For most processes, the conventional and package-free models use identic, generic
data from Ecoinvent in different amounts. Within these processes, a qualitative
analysis of toxicity would be too complex to perform for this study.
The main difference between the systems is the washing procedure at the package-
free retail store. The store uses a dishwasher detergent marked with Ecolabel
(European Commission, 2014). If a detergent complies with these criteria, it restricts
both chemical usage in the product itself as well as in the production phase.
Furthermore, an upper limit exists for how much chemicals can be used per wash
cycle. Because of this, it is unlikely that the package-free model includes toxins that
would make it less desirable than the conventional model.

5.6 Sensitivity analysis

Conventional oat system

As shown in Figure 14, the secondary packaging is a larger contributor in several

impact categories compared to the primary packaging. This raises the question if the
results are bound to an unnecessary inefficiency at the studied company and whether
it is applicable for oats distribution in general. But when the amount of packaging
is minimised for primary packaging, producers may have to increase the secondary
packaging to ensure stability during transport.
The contribution relates to the production of the paper bag and the corrugated board
box, with plastic film contributing to a far smaller share. A main contributing factor
is the car transport, and it is dependent on the assumed size of the consumer grocery
bag, as this affects the allocated impact to the functional unit. The extreme case
where a consumer only buys one package of oats, the car transport itself accounts
for between 75%- 99% of the total impact across all categories. Due to the lack of
quality in the underlying inventory, this constitutes key data, as shown in Figure 5.
But since the scope of this project is to assess the difference between two supply
chain models, this value is of less interest as it is not an inherent attribute in either
system. This line of reasoning also apples for the conventional scenario of rapeseed
oil.

55
 Package-free oat system

The three main contributors are the plastic film that is wrapped around the oat sack
pallets, the paper sack itself, the distribution to the retail store and the paper bag
from the French provider. These represent up to 95% of the total impact of this
system. The values are not sensitive to change. A 30% increase for each of these
three parameters results in an overall increase of 26% for the parameter climate
change, and similar values for the rest of the impact categories.

Conventional rapeseed oil system

As for the case of oats, the results heavily depend on the car transport to the
consumer. However, the packaging in the rapeseed oil system is larger than the
impact of oats, which means that the relative impact of the car transport is lower for
rapeseed oil than for oats. This is because the same assumptions have been made
regarding car transport for both systems, and as the impact is higher for rapeseed
oil, the relative impact of the car transport becomes lower.
Across most impact categories, the most important factor is the plastic bottle, with
production and incineration making up between 50%-90% of the total impact across
all impact categories except ozone depletion. Both the inventoried information
about the bottle itself and the underlying datasets are of good quality, both in
representativeness and timeliness.

Package-free rapeseed oil system

As seen in Figure 15, the impact is slightly more evenly distributed between
different processes for the package-free rapeseed oil system. The production of the
plastic container sticks out, but both the corrugated board box for the plastic
container as well as for the glass bottle contributes about 5 % each, and the glass
bottle slightly over 10 %. For these values, the information is directly measured at
the retail store and the rapeseed oil provider, and the generic data is of good quality.
The performance of the system largely depends on how many times the glass bottle
is reused. This is illustrated in Figure 18, showing where the package-free system
reaches break-even, and in Figure 17, where both a scenario without the glass bottle
and one with 50 usage cycles is assumed. No quantitative data could be obtained
about how many times customers reuse the glass bottle, but the sensitivity analysis
has considered many possible scenarios.

56
 5.7 Impact compared to the food production

The percental impact of packaging

In Table 13, a summary of previous life cycle assessments of oats and rapeseed oil
is presented. As the studies partly use different impact assessment methodologies,
it is not possible to directly translate the results from this study to the results
presented in the table since one would be required to perform unit conversion first.
See notations provided in the table.
Climate change

Acidification

- average
Eutrophication

radiation
Ionizing

Ozone deletion

ozone form.
Photochemical

matter
Particulate
Sources
Impact of 1 kg oats

(Nielsen, et al., 2007)c 0,57 0,006b 0,033b 2,2E-04

(Nielsen, et al., 2007)ac 0,39 0,0064b 0,046b 2,5E-04

(Nielsen, et al., 2007) 0,66 0,007b 0,017b 3,0E-04

(Ecoinvent; Mouron, P, 2016)c 0,714 0,093 0,144 0,03 3,4E-08 0,003 0,002

(Tynelius, 2008) 0,74

(Hille, Solli, Refsgaard, Krokann, &
0,522
Berglann, 2012)
Proposed value 0,5 0,093 0,144 0,03 3,4E-08 0,003 0,002
Impact of 1 kg rapeseed oil
(Angervall & Sonesson, 2011) 1,4
(Ecoinvent; Dauriat, A, 2016) 2,194 0,047 0,074 0,105 1,6E-07 0,01 0,002
(Schmidt, 2007) 2,39 b b
0,026 0,211 1,5E-07 8,7E-04

(Schmidt, 2007)d 2,17 0,024b 0,166b 1,4E-07 8,2E-04

d 2,79 b b
(Schmidt, 2007) 0,029 0,294 1,6E-07 9,5E-04

(Nielsen, et al., 2007)c 1,51 0,012b 0,149b 3,7E-04

(Nielsen, et al., 2007)ac 0,95 0,011b 0,181b 4,5E-04

Proposed value 1,3 0,047 0,074 0,105 1,5E-07 0,01 0,002

a) This refers to organic farming. b) These values are computed with a different LCIA methodology
than used in this report. c) These scenarios do not include the refinery process, e.g. turning oats into
flakes or rape seed into oil. d) These cases present results from the sensitivity analysis.
Table 13: Overview of previous life cycle assessments of oats and rapeseed oil
57
Both the oats and the rapeseed oil in this study are produced from organic farming,
and therefore this will be assumed here as well. The results for the organic farming
are not as comprehensive as those for conventional models, so an average must be
assumed. For photochemical ozone formation, different sources all employ a similar
LCIA model. However, as seen in Table 13, the difference between the value
computed by Ecoinvent greatly differs from the other values, which must be due to
an underlying methodological inconsistency. Attempts were made to correct for
this but without success, which is why the value from Ecoinvent was chosen.

Figure 21: Comparison between packaging and food impact

In Figure 21, the impact of packaging is compared to the impact of the food product
itself. The proposed values from Table 13 have been corrected with all transports
that affects the food product in all different system models, e.g. modelling the
transport of one kilo food product from gate to recycling.
Each bar represents a different scenario, including all baselines scenarios and all
theoretical scenarios. The conventional scenarios used here are modelled without

58
the final car transport to the consumer. For the impact category climate change, we
see for instance that the production and distribution of the packaging material itself
constitutes 10 % of the total impact of the food distribution for conventionally
distributed oats, e.g. the production and distribution of the food product itself
accounts for the remaining 90 %. Each impact category has 8 bars, each representing
the different scenarios denoted below. The first four bars correspond to the oat
scenarios, and the final four to the rapeseed oil scenarios.

Food waste generation

Figure 22: Overview of point of break-even between conventional and zero-waste setting zero-
waste food generation as a variable

In Figure 22, a theoretical calculation is presented, where food waste was set as a
variable, satisfying the equation where the conventional and package-free scenario
reach break-even. As discussed earlier, it is very difficult to translate these results
directly to which system is favourable. The tolerance for some categories are far
higher than for others, and no methodology exists to compare them. For several
categories, including acidification and eutrophication, the tolerance is almost
negligible. This means that whichever system best aids in minimising food waste
will be preferable to the other.

Possible areas for differences in food waste generation

Towards the end of the packaging process for oats, the paper bag goes through a
metal detector to avoid any harmful items from reaching the end consumer. If metal
59
is detected, the oats are disposed of. Due to the larger packaging used in the package-
free system, food waste most likely increases as more oats are disposed of when
metal is detected. At the same time, the smaller paper bags used in the conventional
model may increase food waste in the distribution as the larger packaging uses more
robust packaging paper, according to the employees working with warehousing and
distribution.
At the studied package-free retail store, oats are first refilled from the paper bag into
the gravity dispensers, which could amount in smaller losses. Then, the oats are
again refilled into the final consumer packaging. If the consumer fills carefully, no
waste should occur at this stage, but it is a sensitive system as some product easily
ends up on the floor. If we disregard product damage during distribution, food waste
for dry products is most likely slightly higher when using gravity bin dispensers
than conventional, disposable packaging. This issue could be alleviated through
packaging innovation, designing larger packaging tailor-made to fit into gravity
dispensers, thereby eliminating the need for repackaging.
For the oil, no difference during production was observed between larger or smaller
plastic containers. In the store, the oil is stored at ground level and pumped up
through a dispenser. It is unclear which level of utilisation is obtained with this
model, e.g. how empty the plastic container is once emptied. This problem might
multiply as the oil is repackaged several times. As some oil remains in the packaging
even after emptying, repackaging it several times may increase this amount, as some
oil remains in every container.

5.8 Selected survey results

A few chosen results from the survey are displayed here. For a full summary, refer
to Appendix B.

Figure 23: Overview of Likert questions used.

60
 Figure 24: Overview of product categories offered by the retail stores

5.9 Analysis of survey results

Despite efforts to zero in on package free stores, five out of sixteen stores did not
identify themselves with a concept where packaging is entirely excluded, but rather
that they mostly offer package free foods with few items packaged.
In general, the results show that the package-free concept is not unified, and stores
differ both in terms of ideology and how their respective store operates.

Store operation procedure

The concept found at the Danish zero-waste store is not applicable for the package-
free/zero-waste concept at large. Different answers are given to almost every
question.
The deposit system analysed in this study only applies to eight stores, even though
fifteen stores offer oils and vinegars. This is not the case for oats, where all
respondents answered that they use gravity dispensers.
The supply chain model differs between different stores. In this study, it has been
assumed that the store buys food from a wholesale agent. This model cannot be
generalised as seven stores state that they buy food directly from the farm, and six
stores employ a combination between wholesalers and farmers. In terms of transport
packaging, the majority state that they buy most or all food in reusable packaging
from their suppliers, unlike the disposable packaging solution found in this model.
The food product itself also differs, as some stores only sell organic produce,
whereas others also sell conventionally grown produce.
Regarding the conflict between food and packaging waste, stores have different
opinions whether to use packaging to increase shelf life. Stating that the packaging
61
is made from bio-degradable material makes the stores more positive towards it.
This raises the question if the store concept is based on avoiding inorganic materials
such as plastics, but are not ideologically opposed to bio-based packaging if it
promotes environmental benefits.
Finally, to solve the issue with food waste of animal products, one store states that
they only sell dried or salted products of fish and meat.

Values and ideology

The most preferred label for the store concept was ‘unpackaged store’, but different
stores preferred different labels and it does not seem as if they are unified under a
common label. In general, no strong correlations or clear patterns could be identified
when studying the individual responses.
For some stores, it is important to remain vegetarian, whereas others offer animal
products. Not all stores strive towards only offering organic produce and some aim
at offering most products offered by conventional stores whereas others do not. It
seems as if most stores prioritise locally produced food, as all mostly agreed or
agreed that this is important.

62
 6 Discussion

6.1 Previous work

Nessi, Dolci, Rigamonti & Grosso (2016) looked at other dry products than oats.
They found that for pasta, the waste prevention scenario did not result in any major
savings. For breakfast cereals and rice on the other hand, larger savings were found.
In this study, savings were found for the theoretical package-free baseline in the
order of 40-60 %. This can be compared to the savings found by previously
mentioned authors, where waste prevention scenarios for breakfast cereals lead to
net savings of 60-80 %. This difference likely exists because the breakfast cereals
are packed in a plastic bag inside a carton box, meaning that more, and more
resource demanding packaging is required compared to a single paper bag analysed
in this study.
For rapeseed oil, the results from this study are not as strong as those found by
Cleary (2013), where refillable glass bottles are evaluated in Toronto. Cleary
assumed 15 usage cycles of the refillable glass bottles, making the results from this
study comparable. Here, we found potential savings for most impact categories
around 20-30 %, whereas Cleary calculated net savings of about 80 %. This can be
explained as Cleary compared a refillable glass bottle to a conventional single use
glass container, instead of a single use plastic container evaluated here.
As we saw in Table 7, as journey distances increase, single use containers tend to
be preferable to reusable alternatives (Capuz Rizo, 2005; Vogtländer, 2004).
Previous studies showed that return systems perform well when journey distances
do not exceed around 200 kilometres. In this study, we found that for most impact
categories, journey distances that exceed 400 kilometres renders a single-use system
preferable compared to reusable solutions for almost every impact category, and for
some categories even at distances of 100 kilometres. This relates directly to the
weight of the container used for the reusable distribution. If the glass bottle in this
study would be replaced with a reusable, plastic solution, one could expect a higher
tolerance for farther journeys between the store and the washing facility, but this
has not been analysed.
Pagani, Vittuari & Falasconi (2015) also concluded that reducing packaging for pre-
packed salad would lead to large environmental benefits. However, as mentioned

63
earlier, none of these studies account for the conflict between packaging and food
waste. In this study, the relative impact between packaging and food production is
illustrated and the emission savings are put in context. The results show that even
for rolled oats having a relatively small impact across many categories, the relative
impact of packaging is small. If we were to extrapolate these results to other dry
products having higher emissions, the tolerance for food waste would be far smaller.
If the results from this study are applicable to rice, studied by Nessi, Dolci,
Rigamonti & Grosso (2016), even one or two percent increase in food waste would
make the conventional model preferable (Kasmaprapruet, Paengjuntuek, Saikhwan,
& Phungrassami, 2009).

6.2 A zero-waste retail store framework

The survey, despite receiving few responses, shows that the package-free concept is
used in different ways by different stores. This is in line with the results found by
Beitzen-Heineke, Balta-Ozkan & Reefke (2016), where seven interviews with store
owners were conducted.
In chapter 3, the existing base of knowledge was presented. We could see that no
framework or conclusion could be valid for all different products, and many results
depend on the system boundaries, geographical location, how the authors award
credit of substitution and which heterogeneity exists for similar product categories.
This limitation also applies to this study, and a framework setting out to provide an
answer to which distribution model minimises waste generation for all products is
not feasible. However, throughout the literature, a few overarching results have been
identified and have been translated into a framework, hoping to give general
guidance to practitioners in mitigating waste and minimising climate impact.

64
 The framework

Figure 25: Proposed framework for zero-waste in the food industry context. No strict
hierarchy exists, neither between the three areas or the properties within.
By combining the results from the literature review and a few results from the
survey, the framework presented in Figure 25 is recommended as a standard for
the zero-waste retail stores.
The framework consists of three categories: product, distribution, and packaging,
and relates to how both avoid waste and minimise emissions for each of these
three areas. Within each of these categories, different properties have been
identified. For each category, no strict hierarchy exists between the different
properties, yet an attempt has been made to order them in terms of importance. If
we look at packaging, being emptiable and protective, which mitigates food waste,
is arguably more important than the packaging being portable and optimised for a
return-based distribution model. The same line of reasoning applies to the product,
that a durable product that does not generate food waste is more important than an
efficient handling of food surplus that should not have existed to begin with.
However, this order does not always make sense. If we look at rapeseed oil, there
is a large potential for the remaining oil cake to be used for human consumption
instead of going to animal feed. As oil has a long durability, increasing the level of
utilization may be of higher importance for this product.
65
This framework does not account for customer preferences in any extent, and is
only intended to use as an instrument to achieve zero-waste. By applying this
framework, several products usually found in retail stores would most likely be
deemed inappropriate for a zero-waste distribution model. Some examples would
be sensitive, fresh produce such as fresh berries or leafy greens, fresh fish, and
many types of meat with short durability such as chicken or minced meat. It also
does not account for customer convenience in terms of packaging when phasing
out disposable packaging solutions. The results should therefore not be seen as an
attempt to propose an optimal framework for the industry at large, as it excludes
previously mentioned aspects. Again, the purpose of the framework is to be used
by a practitioner of zero-waste, and the framework therefore promotes zero waste,
regardless of convenience sacrifices.
A more in-depth description of each property of the framework and which parts of
the theory section is used as a foundation follows below.

The product

6.2.2.1 Durable products

In line with the definition for zero waste in Table 3, creating a zero-waste framework
for the food sector means that waste is avoided, not minimised. Based on the
discussion in chapter 3.3.2, many products generate excessive waste despite efforts
to enhance packaging design. Instead of these sub-optimised solutions, the zero-
waste should in principle avoid waste by redirecting food production towards more
durable processing. Some examples could be to prioritise salted or dried meat,
fermented milk products or durable vegetables such as starchy roots. A liberal
interpretation, in case changing the inherent property is not an option, is to properly
evaluate the different distribution options for the food product itself. An example is
minced meat which would benefit from being sold frozen instead of fresh. Instead
of selling a waste-generating product, the industry should design products to last
and minimise waste through their inert properties instead of mitigating damage with
packaging innovation.

6.2.2.2 Utilization of food production

In chapter 3.1.2, eco-effectiveness and the importance of upscaling products was
discussed. For food products, this means both avoiding unnecessary waste and to
capture as much (environmental) value of each product as possible. Some examples
include waste from the disposal of unaesthetically looking produce, or by upcycling
bi-products. An example of upcycling is for rapeseed oil production. The rapeseed
cake that remains is high in nutrients and viable for human consumption, but is today
mostly used as animal fodder. The grade of utilization is far greater if the rapeseed
cake goes directly to human consumption than passing through animals, and this
66
should be considered for the zero-waste practitioner. To achieve this, the suppliers
to zero-waste stores must reverse their value proposition thinking. Instead of
farming to create a certain product - such as an orange, straight carrot – the farmer
should consider what he or she grows and which portfolio of products would
optimise the utilization level of the entire yield. An example could be by offering
fresh, fermented and pureed carrots as the value proposition.

6.2.2.3 Redistributable food

As we saw in the framework developed by Eriksson (2015), it is important to create
food which is redistributable in case it cannot be sold. It is difficult to imagine a
cradle-to-cradle mindset for food products as they are consumed and naturally
downgraded, but the core value of cradle-to-cradle is to minimise downcycling of
products. Therefore, utilisation principles based on biogas recycling should as far as
possible be avoided in favour of economic valorisation, conversion, or donation, as
argued in chapter 3.3.2.2. This means that a trade-off between packaging and
redistributable design must be performed. If donation is used as a food surplus
handling method, using packaging may reduce food waste generation during this
stage.

The distribution

6.2.3.1 Regionalize food and packaging sourcing

Every store responded that prioritising regional food is important to their business
concept. This is also in line with the zero-waste theory, as discussed in chapter 3.1.
By regionalizing the supply, municipalities can better control which waste enters
their waste streams and work towards directing these practises in a more sustainable
direction. As discussed in chapter 3.3.3, transport distance is the critical success
factor for any reverse logistics implementation. By regionalizing the supply, we can
avoid a solution where reversed logistics would result in greater emission than a
single-use model. This assumes a theoretical case where even transport packaging
is reusable, making it sensitive to transport distance. For the retail store studied here,
this parameter is of less importance as they use disposable transport packaging.

6.2.3.2 Utilize optimised load capacity

As discussed in chapter 3.3.2.1, driving back and forth to farms to buy small
quantities is sub-optimised compared to a larger, mass distribution of food. If retail
stores source food locally and solve the logistical challenges by several, small-scale
transports to and from farms, this will result in higher emissions and invalidate the
usefulness of the zero-waste concept. Therefore, the logistics should be
characterised by large vehicles with maximised load factors. This is directly
applicable at larger stores, whereas smaller stores will be bound to resolving this
67
issue with synergistic solutions to combine their demand into larger logistical
systems.

The packaging

6.2.4.1 Emptiable and protective

The core purpose of packaging is to protect the product, and the agreement for this
is strong in all literature, including the zero-waste theories. Packaging should protect
the product and minimise any spoilage during storage and transport.
Another key aspect is to design emptiable packages. Because the zero-waste model
includes more packaging stages between producers and consumers, it is vital that no
additional waste occurs in these transition stages. The packaging should therefore
be designed to make it entirely emptiable. An example is the refill solution
employed at the Danish store studied here. As customers refill their own packaging
from a larger, plastic container, it’s vital that all rapeseed oil is emptied. As
discussed in chapter 5.7.2, a theoretical break-even point can be calculated between
the conventional and zero-waste case where the conventional is preferable due to
the increase in food waste for the zero-waste model. By designing packaging that is
entirely emptiable, this problem can be mitigated.

6.2.4.2 Reusable with material harmonisation

Eleven stores identified themselves with the ambition to entirely free their store
from disposable packaging, while 5 claimed they also offer a few items in disposable
packaging. As discussed in section 3.1, material should be downcycled as little as
possible, which is not the case of disposable packaging. The containers should be
designed to last for as long as possible. But inevitably, these containers will
eventually end up in material recycling. As shown in chapter 3.3.2.2, mixing
packaging material decreases the grade of recyclability. To overcome this issue,
packaging should be homogenous in terms of material choice, so that once it is sent
to recycling, it can be recovered through closed-loop recycling, per Table 6.

6.2.4.3 Portable
Lastly, the packaging should be portable. This means that it is weight efficient and
minimises emissions from transport. As discussed in chapter 3.3.3, a zero-waste
concept will be characterised by reverse logistics. This means that packaging should
not only be light, but also portable in terms of the opposite transport route. Some
examples of this could be to design foldable packaging, which once used can be
transported back to the supplier at a low cost.

68
 Package free or zero-waste

It is unclear how to interpret the term ‘package free’, as no previous literature

addresses this. If a store reduces packaging without considering all aspects covered
in the zero-waste framework in Figure 25, emissions will likely be higher than with
a single-use packaging model. However, it is not known how closely related stores
identifying themselves as package free are to the zero-waste framework presented
here. The zero-waste framework presented above is based on many previous articles
and provides a model that standardises the store concept and ensures reduced
emissions. If the stores identify themselves with this framework, it would reduce
ambiguity and make it easier for customers to know what they are buying into if it
was adopted.

6.3 Combining models

Answering how package-free stores operate around the world is not possible due to
the few responses from the survey. Little can be said about which positive and
negative aspects exist when stores adopt this model, as only a few earlier studies
have been found, and that the results from this study are not translatable to other
store concepts. The discussion that follows will therefore relate to how retail stores
can benefit from combining the zero-waste approach with a conventional model.

Impact results

Under the condition that food waste is not increased, the potential savings for oats
exist. The packaging constitutes roughly 5-10% of the total impact of the product,
and few indicators of an increase in food waste have been identified. If we account
for a change in user behaviour towards employing reusable packaging, the net
savings for the product itself is close to 9%. Based on the review of previous studies
in Table 5, the results are consistent with several different dried products, and
probably consistent for most dry food offered in conventional stores. The savings
on a store level are therefore significant. However, the tolerance for food waste is
most likely much lower for products with a higher impact profile, which makes this
model more suitable for products with a low environmental burden.
It is not known if a package-free or a conventional distribution model leads to less
food waste. If no indicators exist that the food waste would increase by adopting a
system with gravity dispensers, it would be sensible for the industry to adopt it and
redesign their current packaging model as it at least can be concluded that emissions
are reduced from packaging. This conclusion is restricted to the environmental
69
perspective and fails to regard the economic incentive. As single-use packaging is
the dominant solution found on the market, a ‘bias-of-the-present’ exists,
transitioning the burden of proof to the zero-waste model. This is because the
economic cost of a transition must be motivated by an actual increase in
environmental performance to be motivated. In other words, had no existing system
been in place, food waste generation for both systems been unknown and no
indicators that a conventional model would generate less waste than a package-free
model existed, the package-free model would be preferable.
For liquids, the results are more inconclusive. Firstly, the model is sensitive to glass
bottle reuse, and if customers do not properly return glass bottles to be washed, the
system results in large net losses in environmental performance. Secondly, if larger
retail stores adopt the cleaning model, they will likely implement a centralised
solution. This is then sensitive to the distance to the washing facility. If the one-way
transport distance from store to washing facility exceeds 300-400 kilometres, it is
better to keep the existing packaging solution. Lastly, due to the viscosity of oil,
several other product categories are more suitable for a refill solution as the
packaging will be easier to clean and better emptied. If a store were to adopt this, it
is recommended to start with thin liquids such as vinegar and evaluate the
performance before trying to implement it for thicker fluids as oil.
The survey illustrated that several stores offer milk products, fish, and meat, which
is consistent with earlier findings discussed in chapter 3.4. Also, discussed in the
theory chapter is how packaging can mitigate food waste for high-impact products
such as meat. Many previous studies show that through improved packaging
technology, the amount of food waste for meat can be reduced, and the emission
savings from this are far greater than what can be achieved by reducing packaging,
as it constitutes a very minor share of the total impact. By restricting the
transitioning towards package-free products to densely packed, dry products with
low environmental impacts and keeping packaging for other product categories, this
conflict can be overcome.

Pragmatic evaluation of relative packaging impact

A common misconception of an optimised distribution of food is to confuse relative

impact to the total impact of the distribution model instead of comparing different
improvements’ relative weight to each other. It is therefore misleading to state that
because packaging constitutes a small share of the total impact of the system, we
can disregard potential savings from reducing the packaging amount. This is
because we are comparing what can be improved with what cannot, creating a biased
foundation for decision-making. Instead, we can find out the relative impact of
packaging by summarising all potential savings in the entire life cycle of the food

70
product, excluding all impacts that cannot be reduced, or only to a small degree to
a very large cost.
What does this mean in practise? If we consider the results from Figure 21, the
relative impact for oats in the conventional scenario is somewhere between 0.5%-
25%, depending on impact category. These numbers do not give us the actual room
of improvement for this system. Reasonably, we cannot eliminate impact from
packaging and distribution entirely, as the oats need to be produced, packaged in
something, and then reach the final consumer. We should instead – as we’ve done
in this study – investigate which options exist and how much they reduce the impact.
If we look at our best-case theoretical scenario we saw that we could reduce many
impact categories by about 80%. This number must then be put in relation to which
improvements exist in terms of production and food waste mitigation. Just as with
packaging, we cannot reasonably eliminate the impact from the production of the
food product itself, as it, to some extent, will require fertilisers, tractors, and post-
harvest treatment. We must then look at which options exist to minimise these. This
could be for example switching to biodiesel in the tractors. If this would, as an
example, reduce the impact of climate change by 5%, we need to evaluate the
relative packaging impact as an 80% reduction of packaging compared to a 5%
reduction in production. Going back to our case with conventional distribution of
oats, looking at relative impact between the entire impact of food production and
the entire impact of packaging, the contribution of packaging is 8% for climate
change. If we instead compare the two improvements discussed above, the relative
impact of packaging would be 43%.
The example above is trivial, but is only intended to illustrate the purpose of justly
calculating relative impact. In a real-world scenario, we would compare a packaging
reduction to several other reduction measures, both in terms of production,
distribution, minimising food waste and better utilising food surplus.

71
 7 Conclusions

7.1 Research question 1

RQ1: Which definitions of ‘zero waste’ currently exist within food distribution; are
they contradicting each other, and can a general definition be found which can be
used as a framework for the practitioners?

Through the literature review, the different definitions of zero waste have been
presented and their contradictability is discussed. Initially, based on the theory in
section 3.1, a way of interpreting zero waste is proposed and is differentiated from
zero landfill and zero incineration. This core concept is the underlying philosophy
of the framework presented in section 6.2. But as discussed, both in the theory
section 3 and about food waste in 5.7, the minimisation of waste within food
distribution is not one-dimensional and limited to packaging, but rather must
encompass the entire production of food, the distribution and potential end-of-life
treatment. To enhance the framework, a review of findings in previous literature in
regards to waste minimisation is presented throughout the theory section. It is shown
that just removing packaging material will not mitigate waste as it instead risks
increasing food waste. Also, if any waste would occur, it is important to upcycle it
and choose end-of-life treatments that best utilise it, such as redistributing food
surplus to charity instead of sending it to biogas plants. The theory section also
covers several other aspects, such as how systems of reusable packaging systems
rely on short distribution distances to be preferable. The framework does not claim
to provide an answer to which model of distribution is optimal for every food
product. It rather provides general advice to the practitioner based on general
situations. It is entirely possible that for a certain product, the critical distance of the
returnable packaging system is much higher than discussed here. But for the general
case, minimising the distance of the return-trip will often help ensuring that the
returnable system does not become suboptimal to a disposable alternative.
In short, the framework guides practitioners in designing a retail-store concept that
minimises waste as much as possible while still admitting that it cannot provide a
perfect answer for every food product and must be accompanied with more in-depth
studies of specific products to give guarantees.

72
 7.2 Research question 2
RQ2: For rolled oats and cold pressed rapeseed oil, what is the difference in
environmental impact between the ‘package-free’ distribution model and the
conventional, packaged distribution?

If we disregard the difference in food waste, the zero-waste distribution model show
significant savings potentials for oats, but more inconclusive results for rapeseed
oil.
The savings for oats are not bound to customers bringing their own packaging. Still,
by bringing their own container, the savings are increased further. For oats, the
impact from packaging can potentially be reduced by 60%-80% across all impact
categories. For rapeseed oil, the potential savings are a bit lower, about 40%-50%.
The results show us that for oats, even when customers do not bring their own
packaging and use disposable packaging provided by the store, the potential savings
are still around 50% for most impact categories. A case is also investigated if the
customer only fills half the bag with oats, leading to a lower material utilisation and
is backed up by observed user behaviour. Despite this, the package-free model is
preferable to the conventional one, except for the impact category particulate matter
where a 10% increase was shown.
For rapeseed oil, several weaknesses have been found, making the package-free
model more questionable. It is highly dependent on customers reusing the glass
bottles. If the glass bottle is not reused, e.g. it is disposed of after one use, the
package-free model results in higher emissions across all impact categories. The
relative impact is the lowest for ionising radiation at 150%, and highest for ozone
depletion at 329%. But more importantly, the tolerance for food waste is very low
for several impact categories, which combined with an uncertainty about how much
oil can be utilised in repackaging stages and how much more is lost due to several
containers being used, makes the conventional model the preferred choice until a
more in-depth analysis of how much food waste is generated by each system is done.
This does not mean that the package-free model is less preferable than the
conventional one, but based on the high uncertainty of the model, conventional
stores are not recommended to replace their current packaging solution as that may
lead to higher emissions. Novel stores with no existing packaging solution of choice
may choose between them and should prioritise the package-free model if they can
monitor how many times the glass bottle is reused and which emptiness can be
achieved in the repackaging stages, and compare their results to the thresholds
presented in this report. If they cannot, the conventional model is likely a better
option.

73
 7.3 Research question 3
RQ3: In terms of environmental performance, which positive and negative aspects
exist when grocery stores adopt a package-free model and in which ways could a
store benefit by combining the different models?

As no reliable data could be obtained about how the zero-waste stores operate at
large, no conclusions could be drawn about which positive and negative effects exist
when grocery stores adopt the concept as it is not known how the stores operate in
general. Based on the few responses received, stores operated in many different
ways and how the Danish store operates is not representable. The stores differ both
in terms of their line of products, as other stores also offer meat, fish, frozen
products, and ready-made meals, which was not the case for this Danish store.
Furthermore, some stores only sourced their products directly from farms and not
via a wholesaler as in this context.
Based on the results of oats and previous studies, retail stores should consider
redesigning their packaging distribution of certain dry products. The focus should
be on densely packed products with a low environmental impact, and especially
those using excessive packaging solutions at present. This means that products such
as rice should be excluded, as even the slightest increase in food waste would make
the savings from reduced packaging redundant. For this transition to be successful,
retail stores should keep a close dialogue with producers and enable a system where
larger packaging are designed to fit into gravity containers in the store, thereby
eliminating the stage of repackaging and by that minimising the risk that food waste
is increased. The possibility to use reusable packaging even at the supplier stage has
not been addressed in this study, but based on the results of rapeseed oil, it is most
likely bound to short return distances and is therefore not suitable for exotic
products.
For rapeseed oil, the results are not strong enough to support a shift to a package-
free model. More longitudinal studies need to be done, both evaluating how many
times customers reuse the glass bottles, and how the shift to a zero-waste model
affects food waste generation.
If zero waste becomes an emerging concept, retail stores may have to position
themselves towards it as customers may demand it. The zero-waste framework
presented in section 6.2 serves as a good start for understanding what zero waste
means in a food context. The framework can be used as a foundation for
communicating the zero-waste values to the consumers, and make them aware of
the trade-offs that exists between achieving zero-waste and sacrificing their
convenience. As we discussed in section 3.3.2, products such as certain exotic fruits
and minced meat leads to large wastage, and by communicating this to consumer,
stores could shift their offers to more sustainable alternatives. In regards to the
74
current debate about package-free distribution of food, the framework can also help
communicate the complex relationship between different areas, and that each
product may benefit most by improvements in other areas than weight reduction of
packaging.

75
 8 References

Aarnio, T., & Hämäläinen, A. (2008). Challenges in packaging waste management in the fast food industry. Resoures
Conservation & Recycling, 52, 612-621.
Althaus, H.-J., Hischier, R., & Osses, M. (2007). Life Cycle Inventories of Chemicals. Dübendorf: Ecoinvent.
Andrée, E., & Schütte, J. (2010). Matavfall från livsmedelsbutiker - En analys av den kommunala hanteringen i
Sverige samt en detaljstudie av förhållandena i Umeå kommun. Umeå Universitet, Umeå.
Angervall, T., & Sonesson, U. (2011). Förenklad metod för klimat-/GWP-beräkningar av livsmedel. SIK.
Baumann, H., & Tillman, A. (2004). The Hitch Hiker's Guide to LCA. Lund: Studentlitteratur.
Beitzen-Heineke, E., Balta-Ozkan, N., & Reefke, H. (2016). The prospects of zero-packaging grocery stores to
improve the social and environmental impacts of the food supply chain. Journal of Cleaner Production, 140,
1528-1541.
Beumer. (2016). Beumer Packaging Technology - Safe, Clean, Profitable. Beckum, Germany: Beumer Packaging
Technology.
Bryman, A. (2006). Integrating quantitative and qualitative research: how is it done? Qualitative research, 97-113.
Burke Johnson, R., Onwuegbuzie, A. J., & Turner, L. A. (2007). Towards a Definition of Mixed Methods Research.
Journal of Mixed Methods Research, 112-133.
Büsser, S., & Jungbluth, N. (2009). The role of flexible packaging in the life cycle of coffee and butter. International
Journal of Life Cycle Assessment, 14, 80-91.
Capcon Ltd. (2014). UK Milk Roll Cage. Retrieved 05 13, 2015, from http://www.capcon.ie/uk
Capuz Rizo, S. (2005). A comparative study of the environmental & economic characteristics of corrugated board
boxes & reusable plastic crates. Godella, Valéncia: ITENE.
Cederberg, C., & Nilsson, B. (2004). Livscykelanalys (LCA) av ekologisk nötköttsproduktion i ranchdrift. SIK -
Institutet för livsmedelsteknik och bioteknik.
Cleary, J. (2013). Life cycle assessments of wine and spirit packaging at the product and the municipal scale: A
Toronto, Canada case study. Journal of Cleaner Production, 44, 143-151.
Cleary, J. (2014). A life cycle assessment of residential waste management and prevention. International Journal of
Life Cycle Assessment, 1607-1622.
Coley, D., Howard, M., & Winter, M. (2009). Local food, food miles and carbon emissions: A comparison of farm
shop and mass distribution approaches. Food policy, 34(2), 150-155.
Compacta. (2015). Tea bag packaging machine. Düsseldorf, Germany: Compacta.
De Monte, M., Padoano, E., & Pozzetto, D. (2005). Alternative coffee packaging: an analysis from a life cycle point
of view. Journal of Food Engineering, 66, 405-411.
Detzel, A., Giegrich, J., Krüger, M., Möhler, S., & Ostermayer, A. (2004). LCA of one way PET bottles & recycled
products. Heidelberg, Germany: IFEU.
Dyllick, T., & Hockerst, K. (2002). Beyond the business case for corporate sustainability. Business Strategy and the
Environment, 11, 130-141.
Ecoinvent; Dauriat, A. (2016). Rape oil mill operation, Europe without Switzerland. Zürich: Ecoinvent.
Ecoinvent; Mouron, P. (2016). Ecoinvent 3.3 dataset documnetation - Oat production. Zürich: Ecoinvent.

76
Edwards-Jones, G., Milà i Canals, L., Hounsome, N., Truninger, M., Koerber, G., Hounsome, B., . . . Jones, D.
(2008). Testing the assertion that 'local food is best': the challenges of an evidence-based approach. Trends
in Food Science & Technology, 19(5), 265-274.
Ellen MacArthur Foundation. (2013). Towards the Circular Economy: Economic and Business Rationale for an
Accelerated Transition.
Elofsson, E. (2015). Arla Sveriges framtida lastbärarstruktur - En fallstudie om kompromissen mellan funktion,
ekonomi, arbetsmiljö, miljö och leanprinciper. Stockholm: KTH.
Eriksson, M. (2015). Supermarket food waste. Swedish University of Agricultural Sciences, Department of Energy
and Technology. Uppsala: SLU Service/Repro.
Eriksson, M. (2016, 10 07). Interview regarding zero packaging stores. (V. Sjölund, Interviewer)
ERM. (2008). Streamlined Life Cycle Assessment of iGPS, Typical Pooled Wooden Pallets and Single-Use Wooden
Pallets, Intelligent Pooling Systems (iGPS) Company LLC. Environmental Resource Management.
Esteves, J., & Pastor, J. (2004). Using a Multimethod Approach to Research Enterprise Systems Implementations.
Electronic Journal of Business Research Methods, 69-81.
Europan Commision. (2011). International Reference Life Cycle Data System (ILCD) Handbook - Recommendations
for Life Cycle Impact Assessment in the European context. Institute for Environment and Sustainability, Joint
Research Centre. Luxemburg: Publications Office of the European Union.
Europan Commission. (2010). Being wise with waste: the EU's approach to waste management. Belgium:
Luxembourg: Publications Office of the European Union.
European Commission . (2010). International Referenc Life Cycle Data System (ILCD) Handbook - General guide
for Life Cycle Assessment - Detailed Guidance. Institute for Environment and Sustainability, Joint Research
Centre. Luxembourg: Publications Office of the European Union.
European Commission. (2000). Reference Document on Best Available Techniques in the Chlor-Alkali
Manufacturing Industry. Seville: Integrated Pollution Prevention and Control (IPPC).
European Commission. (2014). The EU Ecolabel for Detergents and Dishwashers - "The official European label for
Greener Products".
Fawema. (n.d.). FA 217 - Small Bag Packaging Machine. Engelskirchen, Germany: Fawema GmbH.
Franklin Associates. (2008). Building the business case for reuable transport packaging.
Frischknecht, R., Braunschweig, A., Hofstetter, P., & Suter, P. (2000). Human Health Damages due to Ionising
Radiation in Life Cycle Impact Assessment. Review Environmental Impact Assessment, 20(2), 159-189.
Garnier-Laplace, J., Della-Vedova, C., Gilbin, R., Copplestone, D., Hingston, J., & Ciffroy, P. (2006). First derivation
of Predicted-No-Effect Values for freshwater and terrestrial ecosystems exposed to radioactive substances.
Environmental Science & Technology, 40, 6498-6505.
Gaudreault, C. (2012). Methods for open-loop recycling allocation in life cycle assessment and carbon footprint
studies of paper products. Montreal, Quebec: National Council for Air and Stream Improvement.
Genovese, A., Acquaye, A., Figueroa, A., & Lenny Koh, S. (2014). Sustainable supply chain management and the
transition towards a circular economy: Evidence and some applications. Omega, 66, 344-357.
González-Torre, P., Adenso-Díaz, A., & Artiba, H. (2004). Environmental and reverse logistics policies in European
bottling and packaging firms. International journal of production economics, 88, 95-104.
Greco, S., Wilson, A., Spengler, J., & Levy, J. (2007). Spatial patterns of mobile source particulate matter emissions-
to-exposure relationships across the United States. Atmospheric Environment, 1011-1025.
Greene, J. C., & Hall, J. N. (2010). Dialectics and pragmatism - being of consequence. In A. Tashakkori, & C.
Teddlie, Handbook of mixed methods in social & behavioral research (pp. 119-140). Thousand Oaks: Sage
Publications.
Hallin, A., & Blomkvist, P. (2014). Metod för teknologer - examensarbete enligt 4-fasmodellen. Lund:
Studentlitteratur.

77
Hille, J., Solli, C., Refsgaard, K., Krokann, K., & Berglann, H. (2012). Environmental and climate analysis for the
Norwegian agriculture and food sector and assessment of actions. Working paper, Norwegian Agricultural
Economics Research Institute, Oslo.
Hischier, R. (2016, 04 06). Treatment of waste glass, inert material landfill, CH. Zürich, Switzerland.
Hischier, R., Weidema, B., Althaus, H.-J., Bauer, C., Doka, G., Dones, R., . . . Nemecek, T. (2010). Implementation
of Life Cycle Impact Assessment Methods. St. Gallen: Ecoinvent.
Hobart. (n.d.). Profi FX - Technical data. Retrieved 12 09, 2016, from http://www.hobart-
export.com/wExport_en/products/warewashing/dishwasher/front_loading/FX.php
Hopewell, J., Dvorak, R., & Edward, K. (2009). Plastics recycling: challenges and opportunities. Philosophical
Transactions of the Royal Society B: Biological Sciences, 364(1526), 2115-2126.
Höst, M., Regnell, B., & Runeson, P. (2006). Att genomföra examensarbete. Lund: Studentlitteratur.
Ikaros AB. (2006, 09 21). Säkerhetsdatablad - Ikaros Klorfritt Diskmedel.
ISO. (14040:2006). Environmental management - Life cycle assessment - Principles and framework.
ISO. (14044:2006). Environmental management - Life cycle assessment - Requirements and guidelines.
Jick, T. D. (1979). Mixing qualitative and quantitative methods: Triangulation in action. Administrative science
quarterly, 24(4), 602-611.
Johnson, R. B., & Onwuegbuzie, A. J. (2004). Mixed methods research: A research paradigm whose time has come.
Educational researcher, 33(7), 14-26.
Jungbluth, N. (2005). Comparison of the environmental impact of drinking water vs bottled water. Uster,
Switzerland: ESU Services.
Jurgilevich, A., Birge, T., Kentala-Lehtonen, J., Korhonen-Kurki, K., Pietikäinen, J., Saikku, L., & Schösler, H.
(2016). Transition towards Circular Economy in the Food System. Sustainability, 8(1), 69-78.
JVM A/S. (2015, 02 18). Sikkerhetsdatablad - Afspaendingsmiddel.
Kasmaprapruet, S., Paengjuntuek, W., Saikhwan, P., & Phungrassami, H. (2009). Life Cycle Assessment of Milled
Rice Production: Case Study in Thailand. Europan Journal of Scientific Research, 195-203.
Krausz, R. (2012). All for naught? A critical study of zero waste to landfill initiatives. Lincoln University, Department
of Environmental Management, Lincoln.
Lahega. (2013, 12 02). Säkerhetsdatablad - Lahega alkacip 28.
Leigh, S., Jonson, G., & Smith, D. (2006). Retailing Logistics & Fresh Food: Managing Change in the Supply Chain.
Kogan Page: London.
Levova, T. (2016, 07 08). Treatment of scrap steel, inert material landfill, Europe without Switzerland. Zürich,
Switzerland.
Lindh, H. (2016). Sustainable packaging of organic food . Lund: Division of Packaging Logistics.
Livsmedelsverket. (2016, 11 07). Plast. Retrieved from Förpackningar och köksredskap - Plast:
http://www.livsmedelsverket.se/livsmedel-och-innehall/tillagning-hygien-
forpackningar/forpackningar/plast/
Margni, M., Gloria, T., Bare, J., Seppälä, J., Steen, B., Struijs, J., . . . Jolliet, O. (2008). Guidance on how to move
from current practice to recommended practice in Life Cycle Impact Assessment. UNEP/SETAC Life Cycle
Initiative. Multiple cities: Unpublished.
Morgensen, L., Knudsen Trydeman, M., & Hermansen, J. E. (2009). Vores forbrug af fødevarer har stor betydning
for klimaet . Aarhus: Aarhus Universitet.
Mourad, M. (2015). France moves toward a national policy against food waste. Center for the Sociology of
Organizations. Paris: Sciences Po.
Myhre, G., Shindell, D., Bréon, F.-M., Collins, W., Fuglestvedt, J., Huang, J., . . . Zhang, H. (2013). Climate Change
2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the
Intergovernmental Panel on Climate Change. Cambridge: Cambridge University Press.
78
Nessi, S., Dolci, G., Rigamonti, L., & Grosso, M. (2016). Life Cycle Assessment of Waste Prevention in the Delivery
of Pasta, Breakfast Cereals and Rice. Integrated Environmental Assessment and Management, 12(3), 445-
458.
Nielsen, P. H., Nielsen, A. M., Weidema, B. P., Fredriksen, R. H., Dalgaard, R., Halberg, N., & Sciences, F. o. (2007,
03). LCAfood database. Retrieved from Cash crops (Salgsafgrøder): http://gefionau.dk/lcafood/
Pagani, M., Vittuari, M., & Falasconi, L. (2015). Does packaging matter? Energy consumption of pre-packed salads.
British Food Journal, 117(7), 1961-1980.
Paletten Boerse. (2015, 03). Wooden pallets | Euro Pallets. Retrieved from Euro pallet EUR1:
http://www.palettenboerse.com/wp-content/uploads/2015/03/Palettenboerse-Euro-pallet-EUR1.pdf
Paolotti, L., Boggia, A., Castellini, C., Rocchi, L., & Rosati, A. (2016). Combining livestock and tree crops to
improve sustainability in agriculture: a case study using the Life Cycle Asessment (LCA) approach. Journal
of Cleaner Production, 351-363.
Pettersen, M., & Hægermark, W. (2016, 10 02). Å beskytte maten er emballasjens viktigste oppgave. Aftenposten.
Plastics Europe. (2008). Environmental Product Declarations of the European Plastics Manufacturers - Linear low
density polyethylene (LLDPE). Brussels: Plastics Europe.
Pretty, J. (2005). Farm costs and food miles: An assessment of the full cost of the UK weekly food basket. Food
policy, 30(1), 1-19.
PTR . (2002). LCA of potential environmental impacts of Finnish drinks packaging systems. Helsinki: PTR -
Association of Packaging Technology and Research.
Rabl, A., & Spadaro, J. (2004, August). The RiskPoll software. Retrieved from www.arirabl.org
RobotWorx. (n.d.). KUKA. Fairground St. Marion, Ohio: RobotWorx - A Scott Technology Ltd. Company.
Robson, C. (2002). Real World Research. Oxford: Blackwell Publishers.
Rosenbaum, R., Bachmann, T., Gold, L.-S., Huijbregts, M., Jolliet, O., Juraske, R., . . . Hauschild, M. (2008). USEtox
- the UNEP-SETAC toxicity model: recommended characterisation factors for human toxicity and freshwater
ecotoxicity in life cycle impact assessment. International Journal of Life Cycle Assessment, 13, 532-546.
Rowley, J., & Slack, F. (2004). Conducting a literature review. Management research news, 31-39.
Roy, P., Nei, D., Orikasa, T., Xu, Q., Okadome, H., Nakamura, N., & Shiina, T. (2009). A review of life cycle
assessment (LCA) on some food products. Journal of Food Engineering, 90, 1-10.
Rydh, C., Lindahl, M., & Tingström, J. (2002). Livscykelanalys - en metod för miljöbedömning av produkter och
tjänster. Lund: Studentlitteratur.
Sapsford, R. (2007). Survey Research. Thousand Oaks: Sage Publications.
Schmidt, J. (2007). Part 3: Life cycle assessment of rapeseed oil and palm oil. Ph.D thesis, Aalborg University,
Department of Planning and Development.
Silvenius, F., Grönman, K., Katajajuuri, J.-M., Soukka, R., Koivupuro, H.-K., & Virtanen, Y. (2014). The Role of
Household Food Waste in Comparing Environmental Impacts of Packaging Alternatives. Packaging
Technology and Science, 277-292.
Sonesson, U., Cederberg, C., Flysjö, A., & Carlsson, B. (2008). Livscykelanalys (LCA) av svenska ägg (ver.2). SIK
- Institutet för livsmedel och bioteknik.
Sustain. (2008). RTP proves its green credentials. Birmingham: Linpac Allibert.
Svensk Handel. (2009). Svenksarnas Resvanor så reser vi när vi handlar.
Svenska Retursystem. (2016). Svenska Retursystem. Retrieved 11 22, 2016, from http://www.retursystem.se/
Sveriges Bussföretag. (2016). Statistik om bussbranschen.
The Economist. (2016, 12 14). Food packaging is not the enemy of the environment that it is assumed to be. The
Economist.
The Hub's Engineering Pte Ltd. (2011). Roll Container. Retrieved 05 21, 2015, from
http://www.hubs.com.sg/products.php?maincategory=4
79
Transportstyrelsen. (n.d.). Miljözoner. Retrieved 12 09, 2016, from
https://www.transportstyrelsen.se/sv/vagtrafik/Miljo/Miljozoner/
Tynelius, G. (2008). Klimatpåverkan från Axa havregryn i ett LCA-perspektiv. . Lantmännen.
Wallman, M., & Sonesson, U. (2010). Livscykelanalys (LCA) av svensk kalkonproduktion. SIK - Institutet för
livsmedelsteknik och bioteknik.
van Sluisveld, M. A., & Worrell, E. (2013). The paradox of packaging optimization - a characterization of packaging
source reduction in the Netherlands. Resources, Conservation and Recycling, 73, 133-142.
Weidema, B. P., Bauer, C., Hischier, R., Mutel, C., Nemecek, T., Reinhard, J., . . . Wernet, G. (2013). Overview and
methodology - Data quality guideline for the ecoinvent database version 3. St. Gallen: Ecoinvent.
Wikström, F., Williams, H., Verghese, K., & Clune, S. (2014). The influence of packaging attributes on consumer
behaviour in food-packaging life cycle assessment studies - a neglected topic. Journal of Cleaner Production,
73, 100-108.
Williams, H., Wikström, F., Otterbring, T., Löfgren, M., & Gustafsson, A. (2012). Reasons for household food waste
with special attention to packaging. Journal of Cleaner Production, 141-148.
Vogtländer, J. (2004). Corrugated board boxes and plastic container systems: an analysis of costs and eco-costs.
FEFCO.
WRAP. (2010). Single Trip or Reusable Packaging - Considering the Right Choice for the Environment. Banbury.
Väisänen, S., Mikkilä, M., Havukainen, J., Sokka, L., Luoranen, M., & Horttanainen, M. (2016). Using a multi-
method approach for decision-making about a sustainable local distributed energy system: A case study from
Finland. Journal of Cleaner Production, 1330-1338.
Yay, E., & Suna, A. (2015). Application of life cycle assessment (LCA) for municipal solid waste management: a
case study of Sakarya. Journal of Cleaner Production, 284-293.
Yin, R. (2014). Case Study Research: Design and Methods. London: SAGE.
Zaman, A. (2014). Measring waste management performance using the 'Zero Waste Index': the case of Adelaide,
Australia. Journal of Cleaner Production, 66, 407-419.
Zero Waste New Zealand Trust. (n.d.). The end of waste. Retrieved 11 23, 2016, from
http://communityrecyclers.org.nz/
ZWIA. (2009). Zero Waste International Alliance. Retrieved from ZW Definition: http://zwia.org/standards/zw-
definition/
Österbergs. (n.d.). Wrap Aroundmaskin. Göteborg: Österbergs Förpackningsmaskiner.

80
Appendix A

Question Possible answers

Yes / No / We offer mostly
My store aims to remove disposable packaging as far as possible, offering customers bulk food and the
unpackaged food but also sell
possibility to bring their own, reusable packaging.
certain food items packaged.
It is important for our store to prioritize regional produce over imported food.
The vision is to sell only organic produce.
The vision is to sell most products offered in conventional retail stores, such as fish or frozen goods.
It is important that our store is/remains vegetarian, even if we could successfully sell animal products
without packaging. Disagree / Mostly disagree /
Neither agree nor disagree /
We currently buy products in reusable packaging from our food suppliers. Mostly agree / Agree
We're always looking for the largest possible containers from our providers.
If some packaging increases the durability of a product, we would consider using it.
If some packaging increases the durability of a product and the packaging itself is biodegradable (for
instance made out of sugar), we would consider using it.
We work with social integration, such as employing socially disadvantaged people.
We only sell organic products. Yes / No / To some extent
When Fairtrade is available for a product, we always choose it.
Wholesaler / Directly from
Where do you usually buy your food from?
the farm / Other
We offer a deposit system for glass bottles, where customers can bring back empty glass bottles to be
cleaned by us.
Yes / No
We use gravity dispensers for selling dry goods.
We sell dry foods in another way than gravity dispensers, such as large cloth sacks.
Below follow some questions about waste. Fill in which waste procedures take place for each of the
following waste types. Normal waste disposal /
Donation to charity / No
 Transport package from food suppliers. regular waste exists /
 Food waste Recycling / Biogas recycling
 Disposable items, such as hygienic latex gloves for food handling or similar appliances
Dry foods / Oils, vinegar and
other liquids / Soaps and
hygienic products / Cheese or
Which product categories does your store offer? (Multiple) other milk products / Meat
and/or fish / Fresh vegetables
/ Frozen products / Ready-
made meals / Other
Bio-based store / Zero
Below is a list of labels for the package free concept. Which one do you prefer to use for your store’s packaging store / Zero waste
concept? store / Package free store /
Unpackaged store
In your own words, how would you define the store’s concept? Which are the most central aspects that
Free text answer
makes it different from conventional stores?
Europe / North America /
Where is your store located? South America / Asia /
Australia and the Pacific

81
Appendix B

82
Appendix C

Group Ecoinvent name Data source Data quality External references

General inventory
Pallet lifetime and storage
times
Distribution to customer
EUR-flat pallet production Gunnarshögs Gård. Primary, Theoretical (Paletten Boerse, 2015)
Literature
Transport, passenger car, EURO4 Theoretical Theoretical (Svensk Handel, 2009;
Sveriges Bussföretag, 2016)

Oats - conventional
1.1 Plastic film - From Trioplast to Stenqvist
1.1.1 Plastic film – Polyethylene production, linear low Stenqvist Secondary: Saltå
primary density + extrusion plastic film Kvarn
1.1.2 Plastic film – Polyethylene production, linear low Stenqvist Secondary:
tertiary density + extrusion plastic film Gunnarshögs Gård
1.1.3 Lorry distribution Transport, freight, lorry 16-32 metric Stenqvist Primary
ton, EURO4

1.2 Paper bag – From Stenqvist to Saltå

1.2.1 Kraft paper Kraft paper production, unbleached Saltå Kvarn Primary
1.2.2 Paper bag Missing Missing Missing
production process
1.2.3 Corrugated board Corrugated board box production Saltå Kvarn, Stenqvist Primary
box
1.2.4 Lorry distribution Transport, freight, lorry 16-32 metric Saltå Kvarn Primary
ton, EURO4

1.3 Plastic film – From Trioplast to Stora Enso

1.3.1 Plastic film – Polyethylene production, linear low Stora Enso Secondary: Saltå
primary density + extrusion plastic film Kvarn
1.3.2 Plastic film – Polyethylene production, linear low Stora Enso Secondary:
tertiary density + extrusion plastic film Gunnarshögs Gård
1.3.3 Lorry distribution Transport, freight, lorry >32 metric ton, Stora Enso Primary
EURO4

1.4 Corrugated board box – From Stora Enso to Saltå Kvarn

1.4.1 Corrugated board Corrugated board box production Saltå Kvarn Primary
box production
1.4.2Lorry distribution Transport, freight, lorry, 16-32 metric Saltå Kvarn Primary
ton, EURO4

1.5 Plastic film – From Cyklop Plast to Saltå

1.5.1 Plastic film Polyethylene production, linear low Saltå Kvarn Secondary: Saltå
production density + extrusion plastic film Kvarn
1.5.2Plastic film Polyethylene production, linear low Cyklop Plast Primary
production – tertiary density + extrusion plastic film

1.6 Packaging processes – Saltå Kvarn

1.6.1 Machine electricity Electric voltage transformation, from Saltå Kvarn Primary, Theoretical (Beumer, 2016; Fawema;
consumption high to medium voltage, SE Österbergs; RobotWorx)

1.7 Distribution – From Saltå to retailer

1.7.1 Lorry transportation, Transport, freight, lorry 3.5-7.5 metric Saltå Kvarn Primary
to Saltå Warehouse ton, EURO4

83
Group Ecoinvent name Data source Data quality External references
1.7.2 Lorry transportation, Transport, freight, 16-32 metric ton, Saltå Kvarn Primary,
wholesaler EURO4 Assumption
1.7.3 Lorry transportation, Transport, freight, lorry, 7.5-16 metric Saltå Kvarn Primary,
retail ton, EURO4 Assumption

1.8 Transport packaging recycling

1.8.1 Distribution to Transport, freight, lorry, 7.5-16 metric Saltå Kvarn Primary
Fiskeby recycling ton, EURO4
1.8.2 Distribution from Transport, freight, lorry, 7.5-16 metric Saltå Kvarn Primary
Saltå to Högdalens ton, EURO4
incineration
1.8.3 Distribution from Transport, freight, lorry, 7.5-16 metric Saltå Kvarn Primary, Theoretical (Transportstyrelsen)
retail store to recycling ton, EURO4
plant

1.9 Recycling
1.9.1 Lorry transport, to Transport, freight, lorry, 7.5-16 metric Denmark Secondary: Lös
recycling sorting ton, EURO5 Market
1.9.2 Lorry transport, to Transport, freight, lorry >32 metric ton Assumption Assumption
recycling & incineration

Oats – Zero Waste

2.1 Plastic film – From Signode to Lantmännen
2.1.1 Plastic film – Polyethylene production, linear low Lantmännen Secondary: Saltå
primary density + extrusion plastic film Kvarn
2.1.2 Plastic film – Polyethylene production, linear low Lantmännen Secondary:
tertiary density + extrusion plastic film Gunnarshögs Gård
2.1.3 Lorry distribution Transport, freight, lorry, >32 metric Signode Primary
ton, EURO4

2.2 Plastic film – From Trioplast to Stenqvist

2.2.1 Plastic film – Polyethylene production, linear low Stenqvist Secondary: Saltå
primary density + extrusion plastic film Kvarn
2.2.2 Plastic film – Polyethylene production, linear low Stenqvist Secondary:
tertiary density + extrusion plastic film Gunnarshögs Gård
2.2.3 Lorry distribution Transport, freight, lorry, 16-32 metric Stenqvist Primary
ton, EURO4

2.3 Paper sack – From Stenqvist to Lantmännen

2.3.1 Kraft paper Kraft paper production, unbleached Lös Market Primary
2.3.2 Paper bag Missing Missing Missing
production
2.3.3 Corrugated board Corrugated board box production Stenqvist Primary
box
2.3.4 Lorry distribution Transport, freight, lorry, 16-32 metric Lantmännen Primary
ton, EURO4

2.4 Packaging process – Lantmännen

2.4.1 Machine electricity Electric voltage transformation, from Saltå Kvarn Secondary: Saltå (Beumer, 2016; Fawema;
consumption high to medium voltage, SE Kvarn, Theoretical Österbergs; RobotWorx)

2.5 Distribution – From Lantmännen ro retailer

2.5.1 Lorry distribution, Transport, freight, lorry, 16-32 metric GrönFokus Primary
wholesaler ton, EURO4
2.5.2 Lorry distribution, Transport, freight, lorry 3.5-7.5 metric Lös Market Primary
retail ton, EURO4

2.6 Plastic film – From unknown to Smurfit Kappa

Modelled according to similar processes, ref 1.1, 1.3, 1.5, 2.1, 2,2

2.7 Corrugated board box – From Smurfit Kappa to Tapiero

2.7.1 Kraft paper Kraft paper, unbleached Lös Market Primary

84
Group Ecoinvent name Data source Data quality External references
2.7.2 Lorry distribution Transport, freight, lorry 16-32 metric Tapiero Primary
ton, EURO4
2.7.3 Lorry distribution, NTM: Transport, freight, lorry, on Tapiero Primary
on RoRo-ship RoRo-ship

2.8 Transport packaging recycling

2.8.1 Distribution from Transport, freight, lorry, 7.5-16 metric Assumption Secondary: Saltå
Lantmännen to Fiskeby ton, EURO4 Kvarn
2.8.2 Distribution from Transport, freight, lorry, 7.5-16 metric Assumption Secondary: Saltå
Lantmännen to Högdalen ton, EURO4 Kvarn
incineration

2.9 Recycling
2.9.1 Distribution to Transport, freight, lorry, 7.5-16 metric Marius Pederson Primary
recycling sorting ton, EURO5
2.9.2 Distribution to Transport, freight, lorry, >32 metric Marius Pederson Primary
recycling ton, EURO4
2.9.3 Distribution to Transport, freight, lorry, >32 metric Marius Pederson Primary
incineration ton, EURO4

Rapeseed oil – Conventional

3.1 Plastic container – From Terekas to Gunnarshög
3.1.1 Plastic bottle
3.1.2 Plastic film –
primary
3.1.3 Plastic film –
tertiary
3.1.4 Lorry distribution
3.1.5 Lorry distribution,
on RoRo-ship
3.1.6 Lorry distribution
Polyethylene terephthalate production, Gunnarshögs Gård Primary
granulate, bottle grade + injection
moulding
Polyethylene production, high density, Gunnarshögs Gård Primary
granulate + extrusion plastic film
Polyethylene production, linear low Gunnarshögs Gård Primary
density + extrusion plastic film
Transport, freight, lorry, >32 metric ton Gunnarshögs Gård, Primary
Terekas
NTM: transport, freight, lorry, by Terekas Primary
RoRo-ship
Transport, freight, lorry, 3.5-7.5 metric Terekas Primary
ton

3.2 Corrugated board box – From Smurfit Kappa to Gunnarshög

3.2.1 Corrugated board Corrugated board box production Gunnarshögs Gård Primary
box
3.2.2 Fibreboard Fibreboard production, hard Gunnarshögs Gård Primary
3.2.3 Lorry distribution Transport, freight, lorry, 7.5-16 metric Smurfit Kappa Primary
ton, EURO4

3.3 Plastic film – From Bologna to Packteam

3.3.1 Plastic film –
primary
3.3.2 Plastic film –
tertiary
3.3.3 Lorry distribution
Polyethylene production, linear low Gunnarshögs Gård Primary
density + extrusion plastic film
Polyethylene production, linear low Gunnarshögs gård Primary
density + extrusion plastic film
Transport, freight, lorry, 3.5-7.5 metric Packteam Primary
ton, EURO4

3.4 Distribution – From Gunnarshögs Gård to retailer

3.4.1 Lorry transportation, Transport, freight, lorry, >32 metric Gunnarshögs Gård Primary
to wholesaler ton, EURO4
3.4.2 Lorry transportation, Transport, freight, lorry, 7.5-16 metric Assumption Assumption
to retail ton, EURO4

3.5 Transport packaging recycling

3.5.1 Distribution from Transport, freight, lorry, 7.5-16 metric Gunnarshögs Gård Primary
Gunnarshögs Gård to ton, EURO4
Sysav

85
Group Ecoinvent name Data source Data quality External references
3.5.2 Distribution from Transport, freight, 7.5-16 metric ton, Assumption Assumption
retail store to recycling EURO5

3.6 Recycling
Modelled according to the Danish situation for oats, ref 2.9

Rapeseed oil – Zero Waste

4.1 Plastic container – From Terekas to Gunnarshög
Identical to the conventional scenario, ref 3.1-3.2

4.2 Plastic film – From Bologna to Packteam

Identical to the conventional scenario, ref 3.3

4.3 Distribution – From Gunnarshög to Lös Market

4.3.1 Steel cage Reinforced steel production Lös Market Primary
production
4.3.2 Lorry distribution to Transport, freight, lorry, >32 metric Gunnarshögs Gård Primary
GrönFokus ton, EURO4
4.3.3 Lorry distribution, Transport, freight, lorry, 7.5-16 metric Lös Market Primary
to Lös Market ton, EURO5

4.4 Glass bottle washing

4.4.1 Electricity Electricity voltage transformation from Lös Market Primary (Hobart)
consumption high to medium voltage, DK
4.4.2 Sodium hydroxie Chlor-alkali electrolysis, Lös Market Primary (JVM A/S, 2015; Lahega,
mercury/membrane cell 2013; Ikaros AB, 2006;
Althaus, Hischier, & Osses,
2007; European
Commission, 2000)
4.4.3 Potassium Potassium hydroxide production Lös Market Primary (JVM A/S, 2015; Lahega,
hydroxide 2013; Ikaros AB, 2006)

4.5 Plastic film – from DE producer to DE corrugated board box producer

4.5.1 Plastic film – Polyethylene production, linear low Assumption Secondary:
primary & tertiary density + extrusion plastic film Gunnarshögs Gård
4.5.2 Lorry distribution Transport, freight, lorry >32 metric ton, Assumption Assumption
EURO4

4.6 Corrugated board box – From DE producer to DE glass producer

4.6.1 Corrugated board Corrugated board box production Lös Market Primary
box
4.6.2 Lorry distribution Transport, freight, lorry, >32 metric Assumption Assumption
ton, EURO4

4.7 Glass bottle – From DE producer to Glaeserundflaschen

4.7.1 Glass production Packaging glass production, white Lös Market Primary
4.7.2 Lorry distribution Transport, freight, lorry >32 metric ton, Glaeserundflaschen Primary
EURO4

4.8 Glass bottle – from Glaeserundflaschen to Lös Market

4.8.1 Lorry distribution Transport, freight, lorry 7.5-16 metric Glaeserundflaschen Primary
ton EURO4
4.8.2 Lorry distribution, NTM: Transport, freight, lorry, by Glaeserundflaschen Primary
on RoRo-ship RoRo-ship

4.9 Transport packaging recycling & recycling

Modelled according to the conventional rapeseed scenario, ref 3.5-3.6, and the zero-waste scenario for oats, ref 2.9

86